blob_id
stringlengths
40
40
directory_id
stringlengths
40
40
path
stringlengths
2
327
content_id
stringlengths
40
40
detected_licenses
listlengths
0
91
license_type
stringclasses
2 values
repo_name
stringlengths
5
134
snapshot_id
stringlengths
40
40
revision_id
stringlengths
40
40
branch_name
stringclasses
46 values
visit_date
timestamp[us]date
2016-08-02 22:44:29
2023-09-06 08:39:28
revision_date
timestamp[us]date
1977-08-08 00:00:00
2023-09-05 12:13:49
committer_date
timestamp[us]date
1977-08-08 00:00:00
2023-09-05 12:13:49
github_id
int64
19.4k
671M
star_events_count
int64
0
40k
fork_events_count
int64
0
32.4k
gha_license_id
stringclasses
14 values
gha_event_created_at
timestamp[us]date
2012-06-21 16:39:19
2023-09-14 21:52:42
gha_created_at
timestamp[us]date
2008-05-25 01:21:32
2023-06-28 13:19:12
gha_language
stringclasses
60 values
src_encoding
stringclasses
24 values
language
stringclasses
1 value
is_vendor
bool
2 classes
is_generated
bool
2 classes
length_bytes
int64
7
9.18M
extension
stringclasses
20 values
filename
stringlengths
1
141
content
stringlengths
7
9.18M
e5600e211b854f996cc6a0afbd8ff0a854713c5e
e625eedd8ae078dd88a7b1630727bb3a4e9fd76e
/code/plot_power.R
873ebaa71e0cd17623588f697256b495e2613e24
[ "MIT" ]
permissive
yuyouling/Sze_Obesity_mBio_2016
d8c5b746cef6bfc0b028061a002b69b094f1f098
78166b568af1f79dc95cb5e2b12fb45b3cf78197
refs/heads/master
2021-12-22T17:29:26.568788
2017-10-17T12:51:37
2017-10-17T12:51:37
null
0
0
null
null
null
null
UTF-8
R
false
false
2,876
r
plot_power.R
pch <- c(baxter=21, escobar=21, goodrich=21, hmp=21, ross=21, schubert=21, turnbaugh=21, wu=21, zeevi=21, zupancic=21) col <- c(baxter="black", escobar="red", goodrich="green", hmp="blue", ross="orange", schubert="black", turnbaugh="red", wu="green", zeevi="blue", zupancic="orange") bg <- c(baxter="white", escobar="white", goodrich="white", hmp="white", ross="white", schubert="black", turnbaugh="red", wu="green", zeevi="blue", zupancic="orange") names <- c(baxter="Baxter", escobar="Escobar", goodrich="Goodrich", hmp="HMP", ross="Ross", schubert="Schubert", turnbaugh="Turnbaugh", wu="Wu", zeevi="Zeevi", zupancic="Zupancic") build_plots <- function(method){ pred <- read.table(file=paste0("data/process/", method, "_power.predicted"), header=T, stringsAsFactors=FALSE) metrics <- unique(pred$metric) for(m in metrics){ pred_subset <- pred[pred$metric == m,] o <- order(pred_subset$effect_size, pred_subset$study) pred_subset <- pred_subset[o,] effects <- unique(pred_subset$effect_size) n_effects <- length(effects) studies <- unique(pred_subset$study) n_studies <- length(studies) stagger <- seq(-0.3,0.3,length.out=n_studies) tiff_file <- paste0("results/figures/", method, "_", m, "_power.tiff") tiff(file=tiff_file, width=6.0, height=5, units='in', res=300) layout(matrix(c(1,1,3,2,2,3,0,0,0), nrow=3, byrow=T), width=c(1,1,0.4), height=c(1,1,0.2)) par(mar=c(0.5,5,0.5,0.5)) plot(NA, xlim=c(0.7,4.3), ylim=c(0,1), ylab="Power to Detect Effect Size\nWith Original Sampling Effort", xlab="", axes=F) for(e in 1:n_effects){ effect <- pred_subset[pred_subset$effect_size==effects[e],] points(x=e+stagger, y=effect$power, col=col[effect$study], bg=bg[effect$study], lwd=2, pch=pch[effect$study]) } axis(1, at=1:n_effects, labels=FALSE) axis(2, las=2) box() mtext(side=2, at=1.0, line=3, text="A", las=2, font=2, cex=1) abline(v=c(1.5, 2.5, 3.5)) par(mar=c(0.5,5,0.5,0.5)) plot(NA, xlim=c(0.7,4.3), ylim=c(1,max(pred_subset$balanced_n)), ylab="Number of Samples\nNeeded per Group", xlab="", axes=F, log='y') for(e in 1:n_effects){ effect <- pred_subset[pred_subset$effect_size==effects[e],] points(x=e+stagger, y=effect$balanced_n, col=col[effect$study], bg=bg[effect$study], lwd=2, pch=pch[effect$study]) } axis(1, at=1:n_effects, labels=100*effects) axis(2, las=2) box() mtext(side=2, at=1.2*(10^par()$usr[4]), line=3, text="B", las=2, font=2, cex=1) if(method == 'alpha'){ mtext(1, line=2, text = "Effect Size (%)", cex=0.7) } else { mtext(1, line=2, text = "Effect Size (Cohen's d)", cex=0.7) } abline(v=c(1.5, 2.5, 3.5)) par(mar=c(0,0,0,0)) plot(NA, xlim=c(0,1), ylim=c(0,1), axes=F, xlab="", ylab="") legend(x=0.1, y=0.66, legend=names, pch=pch, col=col, pt.bg=bg, pt.cex=1.5, pt.lwd=2) dev.off() } }
2317d75cbbeb11bb33a2b56649faa0f622fafa36
ef572bd2b0515892d1f59a073b8bf99f81d6a734
/man/unPackDataPackName.Rd
938b6de3bd682f28bcd189d2a7021d2a1ba3a391
[ "CC0-1.0" ]
permissive
pepfar-datim/datapackr
5bc604caa1ae001b6c04e1d934c0c613c59df1e6
9275632673e45948db6846513a53c1436cfc0e47
refs/heads/master
2023-08-30T23:26:48.454382
2023-08-11T13:01:57
2023-08-11T13:01:57
170,350,211
9
7
CC0-1.0
2023-09-11T21:53:24
2019-02-12T16:19:47
R
UTF-8
R
false
true
554
rd
unPackDataPackName.Rd
% Generated by roxygen2: do not edit by hand % Please edit documentation in R/unPackDataPackName.R \name{unPackDataPackName} \alias{unPackDataPackName} \title{Extract the name of the datapack} \usage{ unPackDataPackName(submission_path, tool) } \arguments{ \item{submission_path}{Local path to the file to import.} \item{tool}{What type of tool is the submission file? Default is "Data Pack". javascript:;} } \value{ Character vector of the name of the data pack. } \description{ When supplied a submission path, will return the name of the datapack. }
d0e4a01708f3e4faf432ca69400a8ae86c70bf39
2a691a3598dd431eabe4b0f47f6b7df0bf4467ce
/man/cv.softKv.Rd
102c741f30222152a90a773f0ef38280ea980055
[]
no_license
mengchen18/omic3plus
62a26efbf0ae27cadfd1746df74f2602282b0ef9
34766eea3c4cbf99d0e4e05f19a46fbfa0dd5234
refs/heads/master
2021-01-21T13:25:37.709955
2018-01-23T21:50:57
2018-01-23T21:50:57
102,119,183
0
0
null
null
null
null
UTF-8
R
false
true
1,297
rd
cv.softKv.Rd
% Generated by roxygen2: do not edit by hand % Please edit documentation in R/cv.softSVD.R \name{cv.softKv} \alias{cv.softKv} \title{cross-validation for softSVD} \usage{ cv.softKv(x, nf = 1, kv.opt = c(0.3, 0.5, 0.8), wv = 1, wu = 1, pos = FALSE, maxiter = 50, tol = sqrt(.Machine$double.eps), verbose = FALSE, init = c("svd", "average")[2], ncores = 1, fold = 5, nstart = 1, seed = NULL, loorss = FALSE) } \arguments{ \item{x}{input matrix} \item{nf}{number of component} \item{kv.opt}{optional value for sparsity on right singular value} \item{wv}{weight for columns} \item{wu}{weight for rows} \item{pos}{whether retein non-negative results} \item{maxiter}{maximum number of iteration} \item{tol}{convergence tolerance} \item{verbose}{if print the progress} \item{init}{how to initialize the algorithm. if no sparsity, svd is fast.} \item{ncores}{the number of cores used, passed to mclapply} \item{fold}{fold number in cross validation} \item{nstart}{how many time the k-fold cross validation should be done} \item{seed}{set seed} \item{loorss}{if the Leave-one-out procedure should be used in matrix reconstruction} } \description{ This function use k-fold cross-valiation method to optimize the sparsity of right singular values } \seealso{ \code{\link{cv.softSVD}} }
6f2d9aab37ae5b9e7ef82da2e22cb8e259cc7c17
f4f1cd0987ad6c7a71fc400f41ca3f22e3f6ed53
/man/pImport.Rd
6f2bebd95caf2a018722b3ec0a81384b0fc6dd08
[]
no_license
jordandeherrera/prioritizationMatrix
6ee0e89e12e281c111e2794d5565e7cccd3ab0ba
16484b6f3a6a97e90dc3e899b57e9d77004e0f52
refs/heads/master
2022-01-21T03:50:31.217496
2019-06-25T05:55:01
2019-06-25T05:55:01
null
0
0
null
null
null
null
UTF-8
R
false
true
349
rd
pImport.Rd
% Generated by roxygen2: do not edit by hand % Please edit documentation in R/functions.R \docType{data} \name{pImport} \alias{pImport} \title{Sample Priority Values} \format{An object of class \code{numeric} of length 4.} \usage{ pImport } \description{ An example of priority numeric values. } \examples{ \dontrun{ pImport } } \keyword{datasets}
359d60803b2171a85519a43a77542bdfefb42755
df07a558779d1f2da01e1f47bda210a233028035
/tests/testthat/test-features-brolgar.R
16f1d7ef965fb7316a9cdb365c93d94501ccdb04
[ "MIT" ]
permissive
jesserp92/brolgar
1d208ed805eff9c01b7c2291719abf8670bf453e
93d105dce4d88ca21f3035b1688dd2ad0ef78a99
refs/heads/master
2023-07-03T08:18:18.389004
2021-08-11T01:55:56
2021-08-11T01:55:56
null
0
0
null
null
null
null
UTF-8
R
false
false
1,111
r
test-features-brolgar.R
wages_brolgar <- wages %>% features(ln_wages, feat_brolgar) test_that("feat_brolgar returns the right names", { expect_equal(names(wages_brolgar), c("id", "min", "max", "median", "mean", "q25", "q75", "range1", "range2", "range_diff", "sd", "var", "mad", "iqr", "increase", "decrease", "unvary", "diff_min", "diff_q25", "diff_median", "diff_mean", "diff_q75", "diff_max", "diff_var", "diff_sd", "diff_iqr" )) }) test_that("feat_brolgar returns the right dimensions", { expect_equal(dim(wages_brolgar), c(888, 26)) }) library(dplyr) test_that("feat_brolgar returns all ids", { expect_equal(n_distinct(wages_brolgar$id), 888) })
985d6a8ff93d6b850b66c8fdc9d1d188df8bc340
ffdea92d4315e4363dd4ae673a1a6adf82a761b5
/data/genthat_extracted_code/AER/examples/USInvest.Rd.R
14f4fa10cda20eb7e1c72deb22f839af0cce27a0
[]
no_license
surayaaramli/typeRrh
d257ac8905c49123f4ccd4e377ee3dfc84d1636c
66e6996f31961bc8b9aafe1a6a6098327b66bf71
refs/heads/master
2023-05-05T04:05:31.617869
2019-04-25T22:10:06
2019-04-25T22:10:06
null
0
0
null
null
null
null
UTF-8
R
false
false
894
r
USInvest.Rd.R
library(AER) ### Name: USInvest ### Title: US Investment Data ### Aliases: USInvest ### Keywords: datasets ### ** Examples data("USInvest") ## Chapter 3 in Greene (2003) ## transform (and round) data to match Table 3.1 us <- as.data.frame(USInvest) us$invest <- round(0.1 * us$invest/us$price, digits = 3) us$gnp <- round(0.1 * us$gnp/us$price, digits = 3) us$inflation <- c(4.4, round(100 * diff(us$price)/us$price[-15], digits = 2)) us$trend <- 1:15 us <- us[, c(2, 6, 1, 4, 5)] ## p. 22-24 coef(lm(invest ~ trend + gnp, data = us)) coef(lm(invest ~ gnp, data = us)) ## Example 3.1, Table 3.2 cor(us)[1,-1] pcor <- solve(cor(us)) dcor <- 1/sqrt(diag(pcor)) pcor <- (-pcor * (dcor %o% dcor))[1,-1] ## Table 3.4 fm <- lm(invest ~ trend + gnp + interest + inflation, data = us) fm1 <- lm(invest ~ 1, data = us) anova(fm1, fm) ## More examples can be found in: ## help("Greene2003")
262d1b80f537d3ae7772e61d52d1aa6e2cb61ebb
35df119b1c18fe0df6a62f03a40ae0fc20fe11fd
/6-gganimations.R
7c8e478693230061c04b42cebdf323a823311a67
[]
no_license
linchen-deng/IMDB-webscraping
16774c7a436c8f8d6bae88e971c3abecaae06034
0a12f5437ba90c2c064615cafacc23e31c45a575
refs/heads/master
2020-09-20T06:50:19.190257
2020-02-05T22:27:50
2020-02-05T22:27:50
224,403,921
0
0
null
null
null
null
UTF-8
R
false
false
2,901
r
6-gganimations.R
setwd(dirname(rstudioapi::callFun("getActiveDocumentContext")$path)) source("3-get_data.R") # This file generates animations that descibe the change of number of movies by genres throughout years from 1950 - 2020 # =============== First Animation ============================ if (!dir.exists("./plot_data")){ dir.create(file.path(".","plot_data")) } # if it is not saved in the folder, we will run the code again. if (!"animation.RData" %in% dir("plot_data")){ # stack all data frames into one data frame final_table <- data.frame() for (i in 1:length(genre_list)){ final_table <- final_table %>% rbind(year_data_list[[i]] %>% num_movie_per_year_by_genre(genre=genre_list[i])) } theme_set(theme_classic()) Final_table1 <- final_table %>% group_by(years) %>% mutate(rank = min_rank(-movie_peryear)) %>% ungroup() moving_bar <- ggplot(Final_table1,aes(rank, group= genre,fill = as.factor(genre))) + geom_tile(aes(y = movie_peryear/2,height = movie_peryear)) + geom_text(aes(y = 0, label = paste(genre, " ")), vjust = 0.2, hjust = 1) + coord_flip(clip = "off", expand = FALSE) + scale_y_continuous(labels = scales::comma) + scale_x_reverse() + guides(color = FALSE, fill = FALSE) + labs(title='{closest_state}', x = "", y = "Movies per year") + theme(plot.title = element_text(hjust = 0, size = 22), axis.ticks.y = element_blank(), axis.text.y = element_blank(), plot.margin = margin(1,1,1,4, "cm"))+ transition_states(years, transition_length = 4, state_length = 1) + ease_aes('cubic-in-out') moving_animation <- animate(moving_bar, fps = 20, duration = 25, width = 800, height = 600) # ================= second animation ========================== movie_per_year <- Final_table1 %>% plot_ly( x = ~years, y = ~movie_peryear, size = ~ movie_peryear, color = ~genre, frame = ~years, text = ~genre, hoverinfo = "text", type = 'scatter', mode = 'markers' )%>% layout( xaxis = list( type = "log" ) ) # ================== thrid animation ================================== popular <- Final_table1 %>% filter(genre %in% c("Action","History","Romance","War"))%>% arrange(years) trend_of_movies <- ggplot(popular,aes(x = years, y = movie_peryear, color= genre))+ geom_line() + geom_point() + scale_color_viridis(discrete = TRUE) + ggtitle("Trend of movies") + theme_ipsum() + ylab("Movies per year") + transition_reveal(years) # saving animation results for future use save(moving_animation, movie_per_year, trend_of_movies,file="plot_data/animation.RData") } else { cat("Animations are saved already, reading from folder: plot_data/animation.RData \n") }
87e8b52be1317645aaec4c9c0c72c4da27ef8640
4359d75816ac645b6b80e72b75068f1d4ffc5164
/man/plot_lollipop.Rd
808d773a9b48f468796604c3b4c219baab7e45f9
[]
no_license
Changing-Cities-Research-Lab/seattleViz
04e5e3cfad30a57b632614fed310729ebc2b0a7b
fbcb42776e3dbf74153f24d509801d7b5cfb288d
refs/heads/main
2023-04-13T15:55:07.183707
2021-04-12T23:06:48
2021-04-12T23:06:48
337,885,525
0
2
null
2021-02-25T06:54:14
2021-02-10T23:48:54
null
UTF-8
R
false
true
1,223
rd
plot_lollipop.Rd
% Generated by roxygen2: do not edit by hand % Please edit documentation in R/plot_lollipop.R \name{plot_lollipop} \alias{plot_lollipop} \title{Produce lollipop plot by King County HRA.} \usage{ plot_lollipop( data, var, limits, title = NULL, x_title = NULL, scale_type = "numeric", save = F, savename = "plot.png", caption = paste0(frb_acs_caption_splitline, ses_caption) ) } \arguments{ \item{data}{Data with column for variable of interest with "facet" and "facet_col"} \item{var}{Column name of variable of interest.} \item{limits}{Y-axis limits.} \item{title}{Plot title} \item{x_title}{Title to display along x-axis} \item{scale_type}{Y-axis scale type: "numeric" or "percent"} \item{save}{T if user would like to return plot object and save file, F (default) to just return object.} \item{savename}{File name of map for saving.} \item{caption}{Figure caption} } \value{ Lollipop plot of variable by HRA and SES. } \description{ This function takes in data and produces a horizontal lollipop plot by King County HRA The order of these categories can be adjusted by changing the factor levels of the facet variable. Input data needs columns for variable of interest (titled "var") and HRA. }
6c8c7661faab20a0807c3a11b63465da2d9f77aa
ffdea92d4315e4363dd4ae673a1a6adf82a761b5
/data/genthat_extracted_code/NISTunits/examples/NISTdarcyTOmeterSqrd.Rd.R
7ace7367676ebb1a26a9e523c13c15abaa93008d
[]
no_license
surayaaramli/typeRrh
d257ac8905c49123f4ccd4e377ee3dfc84d1636c
66e6996f31961bc8b9aafe1a6a6098327b66bf71
refs/heads/master
2023-05-05T04:05:31.617869
2019-04-25T22:10:06
2019-04-25T22:10:06
null
0
0
null
null
null
null
UTF-8
R
false
false
203
r
NISTdarcyTOmeterSqrd.Rd.R
library(NISTunits) ### Name: NISTdarcyTOmeterSqrd ### Title: Convert darcy 14 to meter squared ### Aliases: NISTdarcyTOmeterSqrd ### Keywords: programming ### ** Examples NISTdarcyTOmeterSqrd(10)
98b88269665ba9ae0810958a81aa6742c5c99d5c
e428429992f8a90c114570f593b205a44cc96f9f
/cod.R
16da288da1f77c7503709ea3d7da01d8cd693f77
[]
no_license
Ashlo/R
b9c8aea5d28b6de1085bc866683c4010755fedf2
01e16f144e49f23946ad63d68c5ccb97588e3120
refs/heads/master
2023-01-30T01:52:03.795495
2020-12-15T12:30:16
2020-12-15T12:30:16
321,362,375
0
0
null
null
null
null
UTF-8
R
false
false
1,172
r
cod.R
# load required libraries library(caTools) library(class) # function definitions clean.data = function(df) { df = df[-1] df$Gender = factor(df$Gender, levels = c('Male', 'Female'), labels = c(1, 2)) return(df) } load.data = function() { # load the data df = read.csv('Social_Network_Ads.csv') # clean the data df = clean.data(df) return(df) } split.data = function(df) { set.seed(12345) result = sample.split(df$Purchased, SplitRatio = 0.8) return (result) } classify.lm = function(df) { model = lm(formula = Purchased ~ ., data = df) return (model) } predict.values = function(model, df) { predictions = predict(model, newdata = df) return (predictions) } evaluate = function(expected, observed) { cm = table(observed, expected) accuracy = sum(diag(cm)) / sum(cm) return (accuracy * 100) } # function calls df = load.data() result = split.data(df) df.train = df[result == T, ] df.test = df[result == F, ] model.lm = classify.lm(df.train) predictions.lm = predict.values(model.lm, df.test) predictions.lm = ifelse(predictions.lm >= 0.5, 1, 0) accuracy.lm = evaluate(predictions.lm, df.test$Purchased) print(accuracy.lm)
541cbd84245155ecf70dd9f94c2c121cef88ecf7
fadd25738df09516aedb88a53579e7e121ad51f4
/R/inspectMeasure.R
4b96c88bd6937dd52a5cb3f53775959a07e893c2
[]
no_license
JangSeonghoon/maintcivil
2630dee5df3512c5f9ea39b71169b590138e4ddc
7a5c61eedfdd4bb3b10f506b8e11aac11b475f30
refs/heads/master
2021-09-06T21:48:32.855105
2018-02-12T05:23:12
2018-02-12T05:23:12
103,596,243
0
0
null
null
null
null
UTF-8
R
false
false
6,945
r
inspectMeasure.R
#' #'inspect_Measure_graph #' #' @param None #' @return caution Tb(show spots to fix and working priorty) devtools::use_package("magrittr") devtools::use_package("stringr") devtools::use_package("dplyr") devtools::use_package("DMwR") devtools::use_package("ggplot2") devtools::use_package("rJava") devtools::use_package("DBI") devtools::use_package("RJDBC") #' @importFrom compiler cmpfun #' @importFrom magrittr %>% #' @importFrom stringr str_extract #' @importFrom stringr str_split #' @importFrom stringr str_c #' @importFrom dplyr filter #' @importFrom dplyr mutate #' @importFrom dplyr left_join #' @importFrom DMwR centralImputation #' @importFrom ggplot2 ggplot #' @importFrom ggplot2 aes #' @importFrom ggplot2 geom_line #' @importFrom ggplot2 geom_abline #' @importFrom ggplot2 theme_bw #' @importFrom ggplot2 labs #' @importFrom ggplot2 ggsave #' @importFrom ggplot2 scale_x_continuous #' @importFrom ggplot2 theme #' @importFrom ggplot2 element_text #' @importFrom RJDBC JDBC #' @importFrom DBI dbConnect #' @importFrom DBI dbSendQuery #' @importFrom DBI dbExecute #' @importFrom DBI dbFetch #' @importFrom DBI dbHasCompleted #' @importFrom DBI dbWriteTable #' @importFrom DBI dbDisconnect #' @export inspect=function(order){ A=cmpfun( function(){ rm(list=ls()) if(Sys.info()['sysname']=="Windows"){ path= paste0( Sys.getenv("CATALINA_HOME"),"/webapps/bigTeam/" ) }else if(Sys.info()['sysname']=="Linux"){ path="/home/jsh/eclipse-workspace/bigTeam/src/main/webapp/" } load(paste0(path,"RData/inspect.RData")) inspect_file=ls()[(length(ls())-3):length(ls())] drv=JDBC("oracle.jdbc.driver.OracleDriver",paste0(path,"driver/ojbc6.jar")) conn=dbConnect(drv,"jdbc:oracle:thin:@localhost:1521:xe","korail150773","0818") rs=dbSendQuery(conn, paste0("select V4,V8,V9 FROM TEMPORARY WHERE V1=",order)) d=dbFetch(rs) rs=dbSendQuery(conn, paste0("select V4 FROM TEMPORARY")) kind=ifelse(d[1,1]=="ALIGNMENT LEFT","ALL10M", ifelse(d[1,1]=="ALIGNMENT RIGHT","ALR10M", ifelse(d[1,1]=="PROFILE LEFT","PRL10M", ifelse(d[1,1]=="PRFILE RIGHT","PRR10M", ifelse(d[1,1]=="PRFILE LEFT","PRL10M", ifelse(d[1,1]=="TWIST 3M","TWIST3M","SUP")) ) ) ) ) except=as.numeric(d[,3]) max=as.numeric(d[,2]) kind_no=ifelse(kind=="GAGE",3, ifelse(kind=="PRL10M",4, ifelse(kind=="PRR10M",5, ifelse(kind=="ALL10M",6, ifelse(kind=="ALR10M",7, ifelse(kind=="SUP",8, ifelse(kind=="TWIST3M",9,0))))))) startD=(except-0.2)*1000 lastD=(except+0.2)*1000 vector=1:((lastD-startD)*4+1) range=startD+0.25*(vector-1) range=round(range,digits=2) inspect=data.frame("LOCATION"=range) i=1;for(i in 1:4){ if(i!=1){inspect1=inspect} inspect=left_join(inspect,eval(parse(text=inspect_file[i]))[,c(1,kind_no)],by="LOCATION") names(inspect)[length(inspect)]=paste0(names(inspect)[length(inspect)],"_", str_extract(inspect_file[i],"[0-9]{6}")) print( paste0( i,"/",4 ) ) } inspect <- centralImputation(inspect) ##################################################################### inspect_2=inspect %>% filter(LOCATION>=startD,LOCATION<=lastD) j=1;for(j in 1:3){ k=5-j memory=1 cor2=10000 i=1;for(i in 1:100){ if(i!=1) {cor2=ifelse(cor1<cor2,cor1,cor2)} range_original=101:(length(inspect_2[,5])-100) range_positive=i:(length(inspect_2[,k])-(201-i)) # cor1=round(cor(inspect[range_original,5],inspect[range_positive,k])^2,digits=4) cor1=sum(abs(inspect_2[range_original,5]-inspect_2[range_positive,k])) range_negative=(100+i):(length(inspect_2[,k])-(101-i)) # cor1_1=round(cor(inspect[range_original,5],inspect[range_negative,k])^2,digits=4) cor1_1=sum(abs(inspect_2[range_original,5]-inspect_2[range_negative,k])) cor1=ifelse(cor1<cor1_1,cor1,cor1_1) if(i!=1&cor1<cor2){ memory=ifelse(cor1<cor1_1,i,i*(-1)) } if(i==99){ i=abs(memory) if(memory>0){ range_positive=i:(length(inspect_2[,k])-(201-i)) inspect_2[,k]=c(rep(0,100),inspect_2[range_positive,k],rep(0,100)) }else if(memory<0){ range_negative=(100+i):(length(inspect_2[,k])-(101-i)) inspect_2[,k]=c(rep(0,100),inspect_2[range_negative,k],rep(0,100)) } }#if print(paste0( "j=",j," i=",i,"/100"," cor=",cor1," ",cor2," memory=",memory )) }#for(i) }#for(j) ##################################################################### inspect_3=inspect_2 %>% filter(LOCATION>=(except-0.02)*1000,LOCATION<=(except+0.02)*1000) a=which(inspect_3[,1]==except*1000) b=ifelse(max<0,which(inspect_3[,5]==min(inspect_3[,5])),which(inspect_3[,5]==max(inspect_3[,5]))) c=a-b absc=abs(c) len=length(inspect_3[,1]) if(c>0){ inspect_3[,1]=c(inspect_3[-(1:absc),1], rep(0,absc)) }else{ inspect_3[,1]=c(rep(0,absc), inspect_3[-((len-absc+1):len),1]) } inspect_3 %>% filter(LOCATION>=(except-0.007)*1000,LOCATION<=(except+0.007)*1000) %>% ggplot() + aes(x=LOCATION) + geom_line(aes(y=eval(parse(text=names(inspect[2])))),color= '#adc2eb') + geom_line(aes(y=eval(parse(text=names(inspect[3])))),color= '#7094db') + geom_line(aes(y=eval(parse(text=names(inspect[4])))),color= '#24478f') + geom_line(aes(y=eval(parse(text=names(inspect[5])))),color= '#e60000') + geom_abline(slope = 0,intercept = 0) + scale_x_continuous(breaks=seq((except-0.007)*1000,(except+0.007)*1000,2)) + theme_bw()+ labs(x="km",y="검측치") + theme(axis.text.x=element_text(size=13, face="bold"), axis.title.x=element_text(size=15, face="bold"), axis.text.y=element_text(size=15, face="bold"), axis.title.y=element_text(size=15, face="bold")) ggsave(paste(path,"html/inspect.jpg"), width=20,height=14,units=c("cm")) print(except) }#fun )#cmpfun A() }
eba78d2fe3b9a177e99edddd862156b86ca95411
214d008a5228752ee7ababb13230bcb44bba308f
/analysis/engagement_Paper_1/colorScheme.R
4071de4baec5e4d2e28be7cdda27d2829af595f0
[]
no_license
Sage-Bionetworks/elevateMS_analysis
6ef9768d32bfbfbf477949e4a08f28e3708f1515
2828d6407044a4dc08305300480a5c1b8c813227
refs/heads/master
2022-12-28T13:29:04.197254
2020-10-08T01:46:08
2020-10-08T01:46:08
95,137,868
0
0
null
null
null
null
UTF-8
R
false
false
136
r
colorScheme.R
install_load("scales", "ggsci") show_col(pal_npg("nrc")(10)) #pal_npg("nrc")(10) COL_MS_PATIENT = '#DC0000FF' COL_CONTROL = '#00A087FF'
e04ba7ef06e731b615ec2a6f972f46bcb263cfaa
e98b541d19a1893cf0507389888a14b4097607e4
/R/LEN.R
c16d869fd534c88153e77d08f70e7a90172a45bb
[]
no_license
njbultman/xlsxFunctions
6cd647aacd140a45207949fe01b596f00bd0490a
0326efb9af26be1efc7f6e65b782f3a6e30bebcf
refs/heads/master
2023-07-06T00:22:15.572100
2021-08-08T14:26:17
2021-08-08T14:26:17
294,152,526
0
0
null
null
null
null
UTF-8
R
false
false
563
r
LEN.R
#' Count Characters in String or Vector #' #' Given a string or vector, count the number of characters for each string. #' #' @return Returns the number of characters in string or vector. #' @author Nick Bultman, \email{njbultman74@@gmail.com}, February 2021 #' @seealso \code{\link{nchar}} #' @keywords count len nchar #' @export #' @examples #' LEN("hi") #' LEN(c("hey", "hi", "hey")) #' LEN(c(1, "hi", "hey", 2)) #' #' @param text String or vector that you would like its characters counted. #' LEN <- function(text) { y <- nchar(text) return(y) }
f796a77a3805259f6e2eada03087742a73c38590
2c3bc3cd0efb2891081f05b98fff8532280e3789
/other_scripts/stan_helpers/run_stan_models.R
6dcfdb698b8cbc04e3ab68b1c91dfe7c085f6f76
[]
no_license
joelewis101/thesis
459600339d9d9e8be10743c30c43be8670d400ed
0d2831b788792efc91a031edc8dab66fb7c87fdf
refs/heads/master
2023-02-18T01:50:37.883192
2020-08-10T10:36:50
2020-08-10T10:36:50
171,511,551
0
0
null
null
null
null
UTF-8
R
false
false
888
r
run_stan_models.R
df_t0 <- read.csv("chapter_9/stan_df.csv") df_t0$ab N <- nrow(df_t0) t <- df_t0$tstop covariates <- as.matrix(dplyr::select(df_t0,abx_start_time, abx_end_time, hosp_start_time, hosp_end_time, prev_abx_stop_time)) ab_flags <- as.matrix(dplyr::select(df_t0,prev_abx_exposure, ab_this_step)) df_t0$p0 <- 0 df_t0$p1 <- 0 df_t0$p0[df_t0$ESBL_start == 0] <- 1 df_t0$p1[df_t0$ESBL_start == 1] <- 1 start_state = as.matrix(dplyr::select(df_t0,p0,p1)) end_state = df_t0$ESBL_stop stan_data <- list(N = N, t = t, covariates = covariates, start_state = start_state, end_state = end_state, ab_flags = ab_flags) stan_model <- "other_scripts/stan_helpers/stan_model_real_data_exp_fn_loglik.stan" #saveRDS(fit,"/Users/joelewis/Documents/PhD/R/PhD/stan/stan_model_real_data_msm_replica.rda" ) fit <- stan(stan_model, data = stan_data, warmup = 500, iter = 1000, chains = 1, cores = 2, thin = 1)
e43f54b9b928f0b5cad1684d6b8efc51fb2bc70e
0e26112fdff7dd8beacd4a51e38014833a5ebdf7
/src/53_fig3_effplots.R
fb355fee08c4a691790320fdecdd25fd35007f5f
[]
no_license
jkbest2/tunabayes
278fd782d92c60432c01fcf125e48e6ad489a3b6
4911af91a7b8d614392b3039dd86026c341a2982
refs/heads/master
2020-07-15T02:57:17.951095
2019-11-14T04:04:12
2019-11-14T04:06:57
205,463,789
0
0
null
null
null
null
UTF-8
R
false
false
2,604
r
53_fig3_effplots.R
source("src/50_makefigs.R") ## Munge data frames for efficiency plots (figure 3) reshape_effs <- function(df_diag) { df_diag %>% select(model_name, adapt_delta, ess_rate, div_total, div, summary) } ## Get standardized data frames for each dynamics specification fullPT_effs <- reshape_effs(fullPT_diagnostics) %>% mutate(dyn = factor("P-T, estimated m", levels = dyn_levels)) fixedPT_effs <- reshape_effs(fixedPT_diagnostics) %>% mutate(dyn = factor("P-T, fixed m", levels = dyn_levels)) Schaefer_effs <- reshape_effs(Schaefer_diagnostics) %>% mutate(dyn = factor("Schaefer", levels = dyn_levels)) ## Put into a single data frame all_effs <- bind_rows(fullPT_effs, fixedPT_effs, Schaefer_effs) %>% mutate(rhat = map(summary, ~ .x$summary[, 10]), min_rhat = map_dbl(rhat, min, na.rm = TRUE), max_rhat = map_dbl(rhat, max, na.rm = TRUE)) ## Set breaks manually over orders of magnitude breaks <- 10^(-5:5) ## Plot it effplot <- all_effs %>% filter(adapt_delta > 0.75) %>% ggplot(aes(x = adapt_delta, y = ess_rate, color = model_name, shape = div, group = model_name )) + geom_line(size = 0.5, alpha = 0.25) + geom_point(data = filter(all_effs, div), shape = 1, size = 1.5, stroke = 0.3) + geom_point(data = filter(all_effs, !div), shape = 19, size = 1.5) + scale_color_manual(values = param_colors) + xlab("Target acceptance rate") + guides(shape = FALSE, color = guide_legend(title = "", nrow = 1L)) + scale_x_continuous(breaks = ad_vals, labels = c(ad_vals[1:4], "", "", ad_vals[7]), minor_breaks = NULL) + scale_y_log10(name = "Effectively independent samples per second", breaks = breaks, labels = breaks, expand = expand_scale()) + coord_cartesian(y = c(1e-2, 1e3)) + facet_wrap(~ dyn) + theme_jkb(base_size = 8) + theme(plot.margin = margin(l = 1, r = 2), plot.title = element_blank(), axis.ticks = element_line(size = 0.2), axis.line = element_line(size = 0.2), strip.text.x = element_text(vjust = 0, margin = margin(b = 3)), legend.position = "bottom", legend.margin = margin(t = -10)) ## Save a TIFF for Word, and a PDF as a high quality vector image for publication ggsave("figs/fig3_effplot.tiff", effplot, width = 6, height = 4) ggsave("figs/fig3_effplot.pdf", effplot, device = cairo_pdf, width = 6, height = 4)
41f21c07a3af6eabdeb90f489b9f19ecd11bdacc
a7fc036b705c291c306b1ac7a288de39fdc49f4b
/man/get8KItems.Rd
7807359bf381302cf388bc782b9ba36374f2cfb8
[]
no_license
cran/edgar
6476c18e82a360a6e2e725adc8edf0db9fb5a9f7
3e620a52a9d9f00080abf4d24e0f654edc4b59a1
refs/heads/master
2023-08-20T22:42:15.300717
2023-08-10T13:20:02
2023-08-10T15:30:49
48,079,412
10
14
null
null
null
null
UTF-8
R
false
true
1,560
rd
get8KItems.Rd
% Generated by roxygen2: do not edit by hand % Please edit documentation in R/get8KItems.R \name{get8KItems} \alias{get8KItems} \title{Retrieves Form 8-K event information} \usage{ get8KItems(cik.no, filing.year, useragent) } \arguments{ \item{cik.no}{vector of CIK(s) in integer format. Suppress leading zeroes from CIKs.} \item{filing.year}{vector of four digit numeric year} \item{useragent}{Should be in the form of "Your Name Contact@domain.com"} } \value{ Function returns dataframe with Form 8-K events information along with CIK number, company name, date of filing, and accession number. } \description{ \code{get8KItems} retrieves Form 8-K event information of firms based on CIK numbers and filing year. } \details{ get8KItems function takes firm CIK(s) and filing year(s) as input parameters from a user and provides information on the Form 8-K triggering events along with the firm filing information. The function searches and imports existing downloaded 8-K filings in the current directory; otherwise it downloads them using \link[edgar]{getFilings} function. It then reads the 8-K filings and parses them to extract events information. According to SEC EDGAR's guidelines a user also needs to declare user agent. } \examples{ \dontrun{ output <- get8KItems(cik.no = 38079, filing.year = 2005, useragent) ## Returns 8-K event information for CIK '38079' filed in year 2005. output <- get8KItems(cik.no = c(1000180,38079), filing.year = c(2005, 2006), useragent) } }
117aecf92a17be4be75f014e65b78bc6a566605b
93a5d4b6be128af228a4846edb4ba16e5f8db363
/County Profiles.R
4720b994cfde7046c9b51b73963113c10393e57d
[]
no_license
jvanloon93/ACCD-Tools-Scripts
8b3fbb6a71f81b32c9bca80d2696bdd282cc3d7c
32a2920b4494b96a449a04d68f274f121af5e1ab
refs/heads/master
2020-05-24T04:17:37.865099
2020-02-13T17:21:52
2020-02-13T17:21:52
187,089,110
0
0
null
null
null
null
UTF-8
R
false
false
1,140
r
County Profiles.R
library(XLConnect) library(XLConnectJars) library(rJava) library(tidycensus) library(blscrapeR) library(tidyverse) library(usmap) ten_county <- as.vector(fips('PA', county = c('Allegheny', 'Armstrong', 'Beaver', 'Butler', 'Fayette','Greene', 'Indiana', 'Lawrence', 'Washington', 'Westmoreland'))) names(ten_county) <- c('Allegheny', 'Armstrong', 'Beaver', 'Butler', 'Fayette','Greene', 'Indiana', 'Lawrence', 'Washington', 'Westmoreland') Census <- County_profile_Census_Pull(ten_county, Estimates_year = 2018, ACS_year = 2017, dataset = "acs5") BLS <- UN_LF_County_Pull(ten_county, 2018) CEW <- PRA_10_County(2018) %>% bind_rows() CEW <- CEW %>% filter(own_code == 0) %>% select(area_fips, annual_avg_emplvl, annual_avg_estabs, avg_annual_pay) row.names(CEW) <- (c('Allegheny', 'Armstrong', 'Beaver', 'Butler', 'Fayette','Greene', 'Indiana', 'Lawrence', 'Washington', 'Westmoreland') ) book <- loadWorkbook("County_Profile.xlsx", create = TRUE) createSheet(book, "Census") createSheet(book, "LAU") createSheet(book, "CEW") writeWorksheet(book, Census, "Census") writeWorksheet(book, BLS, "LAU") writeWorksheet(book, CEW, "CEW")
f7e39574b835575a74e01fd894be939956219ebe
0ce785a7c4c0641319d4223cd2e6a381006a04d9
/man/makeBSseq.Rd
9d65383b6c45a86e0061292ee48aee57c64a178b
[]
no_license
ttriche/biscuiteer
2d5b9b788de8eb44fc34827d3b9195f6e55a9193
9056279c86487c69f5a0aa57ab102c776e1fa9bc
refs/heads/master
2020-03-10T07:13:24.616315
2019-02-27T18:55:50
2019-02-27T18:55:50
129,257,925
1
3
null
2018-08-27T15:29:22
2018-04-12T13:44:39
R
UTF-8
R
false
true
643
rd
makeBSseq.Rd
% Generated by roxygen2: do not edit by hand % Please edit documentation in R/makeBSseq.R \name{makeBSseq} \alias{makeBSseq} \title{make an in-core BSseq object from a biscuit BED} \usage{ makeBSseq(tbl, params, simplify = FALSE, verbose = FALSE) } \arguments{ \item{tbl}{a tibble (from read_tsv) or a data.table (from fread())} \item{params}{parameters (from checkBiscuitBED)} \item{simplify}{simplify sample names by dropping .foo.bar.hg19 & similar} \item{verbose}{be verbose about what is happening? (FALSE)} } \value{ an in-core BSseq object } \description{ make an in-core BSseq object from a biscuit BED } \seealso{ makeBSseq_HDF5 }
95b4f45aa9d7b344759fba9d0c0e36e1865d1c21
2fe4c16e0377a99e198ab04d5c378ca247ae4329
/Rscript/R/mzkit/man/PrecursorType.Match.Rd
365a57089f43699201cc9fdf5250b5e9e6a34168
[ "MIT" ]
permissive
xieguigang/mzkit
1964d28b0fad5f6d44950fdccdd4a70877f75c29
6391304b550f7e4b8bb6097a6fb1c0d3b6785ef1
refs/heads/master
2023-08-31T06:51:55.354166
2023-08-30T08:56:32
2023-08-30T08:56:32
86,005,665
37
11
MIT
2023-03-14T14:18:44
2017-03-23T23:03:07
Visual Basic .NET
UTF-8
R
false
true
1,518
rd
PrecursorType.Match.Rd
% Generated by roxygen2: do not edit by hand % Please edit documentation in R/match.R \name{PrecursorType.Match} \alias{PrecursorType.Match} \title{Match the precursor type} \usage{ PrecursorType.Match( mass, precursorMZ, charge, chargeMode = "+", tolerance = tolerance.deltaMass(0.3), debug.echo = TRUE ) } \arguments{ \item{mass}{Molecular mass} \item{precursorMZ}{Precursor m/z value of the ion.} \item{charge}{The charge value of the ion} \item{tolerance}{Tolerance between two mass value, by default is 0.3 da, if this parameter is a numeric value, then means tolerance by ppm value. There are two pre-defined tolerance function: \enumerate{ \item \code{\link{tolerance.deltaMass}} \item \code{\link{tolerance.ppm}} }} } \description{ Match the precursor type through min ppm value match. } \examples{ mass = 853.33089 PrecursorType.Match(853.33089, 307.432848, charge = 3) # pos "[M+3Na]3+" charge = 3, 307.432848 PrecursorType.Match(853.33089, 1745.624938, charge = 1) # pos "[2M+K]+" charge = 1, 1745.624938 PrecursorType.Match(853.33089, 854.338166, charge = 1) # pos "[M+H]+" charge = 1, 854.338166 PrecursorType.Match(853.33089, 283.436354, charge = -3, chargeMode = "-") # neg "[M-3H]3-" charge = -3, 283.436354 PrecursorType.Match(853.33089, 2560.999946, charge = -1, chargeMode = "-") # neg "[3M-H]-" charge = -1, 2560.999946 PrecursorType.Match(853.33089, 852.323614, charge = -1, chargeMode = "-") # neg "[M-H]-" charge = -1, 852.323614 }
df1385d444dbc6a727975aedfdf7156cedfda425
0af2caa1b58109fbe67ba984e07b2b709f66d598
/R/metro.R
795bf89197948877664b90cc62c1176bf6b92fd2
[ "MIT" ]
permissive
visualizacionuc/tidy-data
8921aa4673da91005e7dfd91891a44774f85c1ab
2e2b78fb1a1cc69c70278017f2a7a96d214a1920
refs/heads/master
2020-03-14T18:29:58.721330
2018-11-15T15:04:21
2018-11-15T15:04:21
131,742,976
1
1
null
null
null
null
UTF-8
R
false
false
2,360
r
metro.R
library(tidyverse) library(lubridate) # url <- "https://github.com/jbkunst/r-material/blob/gh-pages/201710-Visualizacion-en-el-Analisis/data/2015.04_Subidas_paradero_mediahora_web/2015.04_Subidas_paradero_mediahora_web.csv" # lo fome ----------------------------------------------------------------- url <- "https://tinyurl.com/data-metro-scl" path <- "data/2015.04_Subidas_paradero_mediahora_web.csv" data <- read_delim(path, delim = ";") data data <- data %>% filter(!str_detect(paraderosubida, "[0-9]+-[0-9]")) data <- data %>% filter(paraderosubida != "-") data <- data %>% filter(hour(mediahora) > 0) # interesante ------------------------------------------------------------- ggplot(data) + geom_point(aes(subidas_laboral_promedio, mediahora)) ggplot(data) + geom_point(aes(x = mediahora, y = subidas_laboral_promedio)) ggplot(data) + geom_point(aes(x = mediahora, y = subidas_laboral_promedio), alpha = 0.02, size = 2) + geom_smooth(aes(x = mediahora, y = subidas_laboral_promedio)) ggplot(data) + geom_point(aes(x = mediahora, y = subidas_laboral_promedio, color = paraderosubida), alpha = 1, size = 2) + geom_smooth(aes(x = mediahora, y = subidas_laboral_promedio)) + theme(legend.position = "none") # datita ------------------------------------------------------------------ datita <- data %>% filter(paraderosubida == "ALCANTARA") ggplot(datita) + geom_point(aes(x = mediahora, y = subidas_laboral_promedio), alpha = 1, size = 2) ggplot(datita) + geom_line(aes(x = mediahora, y = subidas_laboral_promedio)) datita <- data %>% filter(paraderosubida == "UNIVERSIDAD DE CHILE") ggplot(datita) + geom_line(aes(x = mediahora, y = subidas_laboral_promedio)) # Comparacion ------------------------------------------------------------- est <- c("ALCANTARA", "UNIVERSIDAD DE CHILE", "PAJARITOS", "PLAZA MAIPU", "LA CISTERNA L2", "BELLAS ARTES", "EL GOLF", "ESCUELA MILITAR", "NUBLE", "PLAZA DE PUENTE ALTO") datota <- data %>% filter(paraderosubida %in% est) datota %>% count(paraderosubida) library(viridis) ggplot(datota) + geom_line(aes(x = mediahora, y = subidas_laboral_promedio, color = paraderosubida), size = 3) + scale_color_viridis(discrete = TRUE, option = "B") + facet_wrap( ~ paraderosubida, scales = "free")
2ebbdb578cb28798468760ac01a5c5f55b56573f
fd0ab0f09d3c07f03e0af82bf93875524c44a0e9
/tmp-tests/test-impute3.R
137abfe9bf24b3f7651dab4818465312a7a1a8e9
[]
no_license
privefl/bigsnpr
b05f9e36bcab6d8cc86fb186c37fe94a6425960a
83db98f974b68132a9a3f3ee7ca388159a4c12b5
refs/heads/master
2023-08-02T13:31:18.508294
2023-06-30T12:15:55
2023-06-30T12:15:55
62,644,144
162
47
null
2022-10-12T16:46:15
2016-07-05T14:36:34
R
UTF-8
R
false
false
10,630
r
test-impute3.R
################################################################################ CODE_IMPUTE_LABEL <- c(0, 0.5, 1, rep(NA, 253)) CODE_IMPUTE_PRED <- c(0, 1, 2, NA, 0, 1, 2, rep(NA, 249)) ################################################################################ imputeChr <- function(Gna, ind.chr, alpha, size, p.train, seed) { # reproducibility if (!any(is.na(seed))) set.seed(seed[attr(ind.chr, "chr")]) # init n <- nrow(Gna) m.chr <- length(ind.chr) loss_fun <- function(x, y, t1 = 0.25, t2 = 0.75, lambda = 0) { mean(((x > t1) + (x > t2) - y)^2) + lambda * ((t1 - 0.25)^2 + (t2 - 0.75)^2) } # correlation between SNPs corr <- snp_cor( Gna = Gna, ind.row = 1:n, ind.col = ind.chr, size = size, alpha = alpha, fill.diag = FALSE ) # imputation nbNA <- integer(m.chr) error <- rep(NA_real_, m.chr) num_pred <- rep(NA_integer_, m.chr) for (i in 1:m.chr) { cat(i) X.label <- Gna[, ind.chr[i]] nbNA[i] <- l <- length(indNA <- which(is.na(X.label))) if (l > 0) { indNoNA <- setdiff(1:n, indNA) ind.train <- sort(sample(indNoNA, p.train * length(indNoNA))) ind.val <- setdiff(indNoNA, ind.train) # ind.col <- ind.chr[which(corr[, i] != 0)] ind.col <- which(corr[, i] != 0) num_pred[i] <- length(ind.col) if (length(ind.col) < 5) ind.col <- setdiff(intersect(1:m.chr, -size:size + i), i) ind.col <- ind.chr[ind.col] # xgboost model bst <- xgboost(data = Gna[ind.train, ind.col], label = X.label[ind.train] / 2, objective = "binary:logistic", base_score = mean(X.label[ind.train]) / 2, nrounds = 10, params = list(max_depth = 4, gamma = 1, alpha = 1), nthread = 1, verbose = 0, save_period = NULL) # learn thresholds on training set pred.train <- stats::predict(bst, Gna[ind.train, ind.col]) lambda <- 8 * loss_fun(pred.train, X.label[ind.train]) opt.min <- stats::optim(par = c(0.25, 0.75), fn = function(t) { loss_fun(pred.train, X.label[ind.train], t[[1]], t[[2]], lambda) }) thrs <- `if`(opt.min$convergence == 0, opt.min$par, c(0.25, 0.75)) # error of validation pred.val <- stats::predict(bst, Gna[ind.val, ind.col, drop = FALSE]) pred.val <- rowSums(outer(pred.val, thrs, '>')) error[i] <- mean(pred.val != X.label[ind.val]) # impute pred <- stats::predict(bst, Gna[indNA, ind.col, drop = FALSE]) pred <- rowSums(outer(pred, thrs, '>')) Gna[indNA, ind.chr[i]] <- as.raw(pred + 4) } } data.frame(pNA = nbNA / n, pError = error, num_pred = num_pred) } ################################################################################ #' Fast imputation #' #' Fast imputation algorithm based on local XGBoost models. **This algorithm #' has not been extensively compared with other imputation methods yet.** #' #' @inheritParams bigsnpr-package #' @param alpha Type-I error for testing correlations. Default is `0.02`. #' @param size Number of neighbor SNPs to be possibly included in the model #' imputing this particular SNP. Default is `500`. #' @param p.train Proportion of non missing genotypes that are used for training #' the imputation model while the rest is used to assess the accuracy of #' this imputation model. Default is `0.8`. #' @param seed An integer, for reproducibility. Default doesn't use seeds. #' #' @return A `data.frame` with #' - the proportion of missing values by SNP, #' - the estimated proportion of imputation errors by SNP. #' @export #' #' @import Matrix xgboost #' snp_fastImpute <- function(Gna, infos.chr, alpha = 0.02, size = 500, p.train = 0.8, seed = NA, ncores = 1) { check_args() Gna$code256 <- CODE_IMPUTE_PRED if (!is.na(seed)) seed <- seq_len(max(infos.chr)) + seed args <- as.list(environment()) do.call(what = snp_split, args = c(args, FUN = imputeChr, combine = 'rbind')) } ################################################################################ #' imputeChr2 <- function(Gna, ind.chr, size, p.train, seed) { #' #' # reproducibility #' if (!any(is.na(seed))) set.seed(seed[attr(ind.chr, "chr")]) #' #' # init #' n <- nrow(Gna) #' m.chr <- length(ind.chr) #' #' # imputation #' nbNA <- integer(m.chr) #' error <- rep(NA_real_, m.chr) #' for (i in 1:m.chr) { #' cat(i) #' X.label <- Gna[, ind.chr[i]] #' nbNA[i] <- l <- length(indNA <- which(is.na(X.label))) #' if (l > 0) { #' indNoNA <- setdiff(1:n, indNA) #' ind.train <- sort(sample(indNoNA, p.train * length(indNoNA))) #' ind.val <- setdiff(indNoNA, ind.train) #' #' ind.col <- -size:size + i #' ind.col[ind.col < 1 | ind.col > m.chr | ind.col == i] <- 0L #' X.data <- Gna[, ind.col] #' #' bst <- xgboost( #' data = X.data[ind.train, ], #' label = X.label[ind.train], #' objective = "multi:softmax", #' base_score = mean(X.label[ind.train]), #' nrounds = 10, #' params = list(max_depth = 4, num_class = 3, gamma = 1, alpha = 1), #' nthread = 1, #' verbose = 0, #' save_period = NULL #' ) #' #' # error of validation #' pred.val <- stats::predict(bst, X.data[ind.val, ]) #' error[i] <- mean(pred.val != X.label[ind.val]) #' # impute #' pred <- stats::predict(bst, X.data[indNA, ]) #' Gna[indNA, ind.chr[i]] <- as.raw(pred + 4L) #' } #' } #' #' data.frame(pNA = nbNA / n, pError = error) #' } #' #' ################################################################################ #' #' #' Fast imputation #' #' #' #' Fast imputation algorithm based on local XGBoost models. **This algorithm #' #' has not been extensively compared with other imputation methods yet.** #' #' #' #' @inheritParams bigsnpr-package #' #' @param size Number of neighbor SNPs to be possibly included in the model #' #' imputing this particular SNP. Default is `100`. #' #' @param p.train Proportion of non missing genotypes that are used for training #' #' the imputation model while the rest is used to assess the accuracy of #' #' this imputation model. Default is `0.8`. #' #' @param seed An integer, for reproducibility. Default doesn't use seeds. #' #' #' #' @return A `data.frame` with #' #' - the proportion of missing values by SNP, #' #' - the estimated proportion of imputation errors by SNP. #' #' @export #' #' #' #' @import xgboost #' #' #' snp_fastImpute2 <- function(Gna, infos.chr, #' size = 100, #' p.train = 0.8, #' seed = NA, #' ncores = 1) { #' #' check_args() #' #' Gna$code256 <- CODE_IMPUTE_PRED #' #' if (!is.na(seed)) seed <- seq_len(max(infos.chr)) + seed #' args <- as.list(environment()) #' #' do.call(what = snp_split, args = c(args, FUN = imputeChr2, combine = 'rbind')) #' } #' #' ################################################################################ #' #' imputeChr3 <- function(Gna, ind.chr, alpha, size, p.train, seed) { #' #' # reproducibility #' if (!any(is.na(seed))) set.seed(seed[attr(ind.chr, "chr")]) #' #' # init #' X <- Gna$copy(code = CODE_IMPUTE_PRED) #' n <- nrow(X) #' m.chr <- length(ind.chr) #' #' # correlation between SNPs #' corr <- snp_cor( #' Gna = Gna, #' ind.row = 1:n, #' ind.col = ind.chr, #' size = size, #' alpha = alpha, #' fill.diag = FALSE #' ) #' print(corr) #' #' # imputation #' nbNA <- integer(m.chr) #' error <- rep(NA_real_, m.chr) #' for (i in 1:m.chr) { #' cat(i) #' X.label <- Gna[, ind.chr[i]] #' nbNA[i] <- l <- length(indNA <- which(is.na(X.label))) #' if (l > 0) { #' indNoNA <- setdiff(1:n, indNA) #' ind.train <- sort(sample(indNoNA, p.train * length(indNoNA))) #' ind.val <- setdiff(indNoNA, ind.train) #' #' ind.col <- which(corr[, i] != 0) #' if (length(ind.col) < 5) #' ind.col <- setdiff(intersect(1:m.chr, -size:size + i), i) #' X.data <- X[, ind.chr[ind.col]] #' #' bst <- xgboost( #' data = X.data[ind.train, ], #' label = X.label[ind.train], #' objective = "multi:softmax", #' base_score = mean(X.label[ind.train]), #' nrounds = 10, #' params = list(max_depth = 4, num_class = 3, gamma = 1, alpha = 1), #' nthread = 1, #' verbose = 0, #' save_period = NULL #' ) #' #' # error of validation #' pred.val <- stats::predict(bst, X.data[ind.val, ]) #' error[i] <- mean(pred.val != X.label[ind.val]) #' # impute #' pred <- stats::predict(bst, X.data[indNA, ]) #' Gna[indNA, ind.chr[i]] <- as.raw(pred + 4L) #' } #' } #' #' data.frame(pNA = nbNA / n, pError = error) #' } #' #' ################################################################################ #' #' #' Fast imputation #' #' #' #' Fast imputation algorithm based on local XGBoost models. **This algorithm #' #' has not been extensively compared with other imputation methods yet.** #' #' #' #' @inheritParams bigsnpr-package #' #' @param alpha Type-I error for testing correlations. Default is `0.02`. #' #' @param size Number of neighbor SNPs to be possibly included in the model #' #' imputing this particular SNP. Default is `500`. #' #' @param p.train Proportion of non missing genotypes that are used for training #' #' the imputation model while the rest is used to assess the accuracy of #' #' this imputation model. Default is `0.8`. #' #' @param seed An integer, for reproducibility. Default doesn't use seeds. #' #' #' #' @return A `data.frame` with #' #' - the proportion of missing values by SNP, #' #' - the estimated proportion of imputation errors by SNP. #' #' @export #' #' #' #' @import Matrix xgboost #' #' #' snp_fastImpute3 <- function(Gna, infos.chr, #' alpha = 0.02, #' size = 500, #' p.train = 0.8, #' seed = NA, #' ncores = 1) { #' #' check_args() #' #' if (!is.na(seed)) seed <- seq_len(max(infos.chr)) + seed #' args <- as.list(environment()) #' #' do.call(what = snp_split, args = c(args, FUN = imputeChr3, combine = 'rbind')) #' } #' #' ################################################################################
1b3c8bacb02aa6f8c20672fe5aa7a5cc769a52b9
76d7a0bb142867e3ff689ef5bc4556a57ae2e6d4
/plot4.R
d5e775a5249b33efc3d1bb3686b1f5bec3a4e316
[]
no_license
patrickbigdigit/Exploratory_Data_Analysis
c2e3f19645aef94e4811bb0b73d1279ba8bcb3e9
e682f8b74c43ff172dfdb97b11c5bee69d90452f
refs/heads/master
2021-01-10T05:27:03.573614
2016-01-09T23:44:19
2016-01-09T23:44:19
49,346,129
0
0
null
null
null
null
UTF-8
R
false
false
1,725
r
plot4.R
## R script for plot 3 # Initial file assignment and data retrival. file <- "household_power_consumption.txt" data <- read.table(file, header=TRUE, sep=";", stringsAsFactors=FALSE, dec=".") # subsetting to limit for two days of data s_data <- subset(data,data$Date %in% c("1/2/2007","2/2/2007") ) # converting to all required variables to numeric class. s_data$Global_active_power <- as.numeric(s_data$Global_active_power) s_data$Sub_metering_1 <- as.numeric(s_data$Sub_metering_1) s_data$Sub_metering_2 <- as.numeric(s_data$Sub_metering_2) s_data$Sub_metering_3 <- as.numeric(s_data$Sub_metering_3) s_data$globalReactivePower <- as.numeric(s_data$Global_reactive_power) s_data$voltage <- as.numeric(s_data$Voltage) # combining date and time values and coverting to datetime format to plot the chart against this variable. var_dt_time <- strptime(paste(s_data$Date, s_data$Time, sep=" "), "%d/%m/%Y %H:%M:%S") # Opening a chart for plotting and updating mfrow for 2*2 array png("plot4.png", width=480, height=480) par(mfrow = c(2, 2)) # Ploting all four graphs as per the requirement plot(var_dt_time, s_data$Global_active_power, type="l", xlab="", ylab="Global Active Power") plot(var_dt_time, s_data$voltage, type="l", xlab="datetime", ylab="Voltage") plot(var_dt_time, s_data$Sub_metering_1, type="l", ylab="Energy Submetering", xlab="") lines(var_dt_time, s_data$Sub_metering_2, type="l", col="red") lines(var_dt_time, s_data$Sub_metering_3, type="l", col="blue") legend("topright", c("Sub_metering_1", "Sub_metering_2", "Sub_metering_3"), lty=1, col=c("black", "red", "blue"), bty="n") plot(var_dt_time, s_data$globalReactivePower, type="l", xlab="datetime", ylab="Global_reactive_power") dev.off()
204fd2ee434b9fea793699e85952214be7dff16f
e3453f8c610ceb4b5022be45b1dc145d690c1358
/man/chatty.Rd
5ed66d027275862f6b3d8ab37334e03f3411e1c8
[]
no_license
kongra/koR
2592b69b7525099d2ceeef2686fb4e0f0005f933
48f06fd6832ac4ccaa7c66438f9a5668072e50a4
refs/heads/master
2018-10-08T03:05:41.749217
2018-08-30T08:18:29
2018-08-30T08:18:29
38,744,631
0
0
null
null
null
null
UTF-8
R
false
true
441
rd
chatty.Rd
% Generated by roxygen2: do not edit by hand % Please edit documentation in R/core.R \name{chatty} \alias{chatty} \title{Debug wrapper generator. Returns a diagnostic wrapper around f. Thanks H. Wickham.} \usage{ chatty(f, prefix = "Processing ") } \arguments{ \item{f}{Function to wrap} \item{prefix}{A prefix for the diagnostic message} } \description{ Debug wrapper generator. Returns a diagnostic wrapper around f. Thanks H. Wickham. }
541e9d24f1a97fb2eff25f8e0223c494f91f9db2
ffdea92d4315e4363dd4ae673a1a6adf82a761b5
/data/genthat_extracted_code/Ecfun/examples/camelParse.Rd.R
206db0aa156311f2597a3857264c9b96c2c9f3e0
[]
no_license
surayaaramli/typeRrh
d257ac8905c49123f4ccd4e377ee3dfc84d1636c
66e6996f31961bc8b9aafe1a6a6098327b66bf71
refs/heads/master
2023-05-05T04:05:31.617869
2019-04-25T22:10:06
2019-04-25T22:10:06
null
0
0
null
null
null
null
UTF-8
R
false
false
579
r
camelParse.Rd.R
library(Ecfun) ### Name: camelParse ### Title: Split a character string where a capital letter follows a ### lowercase letter ### Aliases: camelParse ### Keywords: manip ### ** Examples tst <- c('Smith, JohnJohn Smith', 'EducationNational DefenseOther Committee', 'McCain, JohnJohn McCain') tst. <- camelParse(tst) ## Don't show: stopifnot( ## End(Don't show) all.equal(tst., list(c('Smith, John', 'John Smith'), c('Education', 'National Defense', 'Other Committee'), c('McCain, John', 'John McCain') ) ) ## Don't show: ) ## End(Don't show)
82519d99fed039e28a58dc7715eabb4fe2ab18cb
377b3d754053ae109af8cfb5eb282860a7e134f6
/w6/climr/R/fit.climr.R
04c0e4647d52569c631b6d6f6f561995758a7cbc
[]
no_license
iant04128591/STAT40830
df97f205ab685805535db08f774673a88c9b24eb
ad188ea27f7e54e3e92bd061aa0fdad1a855fdf5
refs/heads/master
2020-08-01T08:13:05.023278
2016-12-08T23:45:21
2016-12-08T23:45:21
73,582,215
0
0
null
null
null
null
UTF-8
R
false
false
1,884
r
fit.climr.R
#' Fit basic statistical moels to climate data #' #' @param obj An object of class \code{climr} from \code{\link{load_clim}} #' @param data_type The type of data to be analysed , either yearly, monthly or quarterly #' @param fit_type The type of model required , eitheither linear regression (\code{lm}), loess or smoothing spline (\code{smooth.spline}) #' #' @return Return a list of class \code{climr_fit} which includes the model details as well as the data set and the fit type used #' @seealso \code{\link{load_clim}}, \code{\link{plot.climr_fit}} #' @export #' @importFrom magrittr "extract2" "%$%" #' @importFrom stats "lm" "loess" "smooth.spline" "na.omit" "predict" #' #' @examples #' ans1 = load_clim('SH') #' ans2 = fit(ans1) #' ans3 = fit(ans1, data_type='monthly', fit_type = 'smooth.spline') #' ans4 = fit(ans1, data_type='quarterly', fit_type = 'loess') fit = function(obj, data_type = c('yearly', 'quarterly', 'monthly'), fit_type = c('lm', 'loess', 'smooth.spline')){ UseMethod('fit') } #' @export fit.climr = function(obj, data_type = c('yearly', 'quarterly', 'monthly'), fit_type = c('lm', 'loess', 'smooth.spline')){ #fund out which data set fit_dat = match.arg(data_type) #find fittig method fit_arg = match.arg(fit_type) #find out which bit of data dat_choose = switch(fit_dat, yearly = 1, quarterly = 2, monthly = 3) #Get the dataset to use curr_dat = obj %>% extract2(dat_choose) #fit some models if(fit_arg == 'lm'){ mod = curr_dat %$% lm(temp ~ x) } else if(fit_arg == 'loess'){ mod = curr_dat %$% loess(temp ~ x) } else if(fit_arg == 'smooth.spline'){ mod = curr_dat %$% smooth.spline(x, temp) } print(mod) out = list(model = mod, data = curr_dat, dat_type = fit_dat, fit_type = fit_arg) class(out) = 'climr_fit' invisible(out) }
1cdb144129b7dde8174e7034e449643f343aaaf9
919de4d940af1e2280cd07665ef8a91789767014
/man/split_vec.Rd
f03f30d89603bff22f141fa1527a66758e8f6523
[ "MIT" ]
permissive
dannydaniel/eegUtils
ded837c9fd84b7ce7f146a573bf54f8b6ff8c638
8af2ff8a9e94b550b54d7222304f1799b61d0162
refs/heads/master
2020-07-22T21:40:20.330013
2019-04-27T11:13:27
2019-04-27T11:13:27
207,336,581
1
0
MIT
2019-09-09T15:03:07
2019-09-09T15:03:07
null
UTF-8
R
false
true
520
rd
split_vec.Rd
% Generated by roxygen2: do not edit by hand % Please edit documentation in R/frequency_analysis.R \name{split_vec} \alias{split_vec} \title{Segment data} \usage{ split_vec(vec, seg_length, overlap) } \arguments{ \item{vec}{Data vector to be split up into segments.} \item{seg_length}{Length of segments to be FFT'd (in samples).} \item{overlap}{Overlap between segments (in samples).} } \description{ Split data into segments for Welch PSD. } \author{ Matt Craddock \email{matt@mattcraddock.com} } \keyword{internal}
62cce5e6263552858d9f3c84c2b6c13e3be67e13
ad9bb63b558e5b60048380eeedfff5bd10d6ac91
/utility.R
5b57e8e80ba4655caa9b8a383405f5d2b21c826d
[]
no_license
mccomark21/NBA_Draft_2020
695c3e69eabcfc97a3494748234620eaff2786ff
92170d95b2e20239d023b8a552e40f1065d93303
refs/heads/main
2023-01-08T08:46:38.792054
2020-11-10T04:17:15
2020-11-10T04:17:15
308,236,471
0
0
null
null
null
null
UTF-8
R
false
false
165
r
utility.R
# Loading Screen for charts---- loadingscreen <- function(plot){ shinycssloaders::withSpinner(plot, type = 8, color = "#0072B2", size = 1.5, hide.ui = FALSE) }
7d9e5156476dbe989d7266c7d385102f6e886027
9e8fcdf0d7f018f540dde7a65d9b3f2d65c9cbc6
/man/feature.missingness.Rd
88887052c77cee0e7c8c1724c1a40c2772ee7582
[ "MIT" ]
permissive
yiheng-aug30/metaboprep
6c02797bd3949c593d8ed70db6e7dcf67cea2eeb
500b051ad8174a54f11b52caa41b9c57f14dcdc4
refs/heads/master
2023-08-29T08:14:57.380002
2021-11-15T13:27:40
2021-11-15T13:27:40
null
0
0
null
null
null
null
UTF-8
R
false
true
1,148
rd
feature.missingness.Rd
% Generated by roxygen2: do not edit by hand % Please edit documentation in R/feature.missingness.R \name{feature.missingness} \alias{feature.missingness} \title{estimate feature missingness} \usage{ feature.missingness( wdata, samplemissingness = NULL, extreme_sample_mis_threshold = 0.5 ) } \arguments{ \item{wdata}{the metabolite data matrix. samples in row, metabolites in columns} \item{samplemissingness}{a vector of sample missingness for each sample} \item{extreme_sample_mis_threshold}{a numeric value above which individuals with sample missingness should be excluded from the feature missingess estimator. Default is 0.5.} } \value{ a data frame of percent missingness for each feature } \description{ This function estimates feature missingess, with a step to exclude poor samples identified as those with a sample missingness greater than 50%. } \examples{ ex_data = sapply(1:5, function(x){rnorm(10, 45, 2)}) ex_data[ sample(1:length(ex_data), 15) ] = NA smis = apply(ex_data, 1, function(x){ sum(is.na(x))/length(x) }) feature.missingness(wdata = ex_data, samplemissingness = smis) } \keyword{feature} \keyword{missingness}
7e9503213f35c99fb4ffdd25cc8fee2587a19f78
9229b21b2365424fe0ff6617cf282b5b03439ef6
/r-demo/helloworld.r
6a287e46d6262a069a2c7ec185a1576a698e828d
[]
no_license
Huize501/GoogleCloud
169b83b0d0324b04dc02ee73b2cd0358f79ed9b1
191c80ed21117a0a0d936916e7d9e618f27377ba
refs/heads/master
2023-02-22T22:01:28.965706
2022-09-16T05:57:51
2022-09-16T05:57:51
116,589,846
4
2
null
2023-02-15T22:49:54
2018-01-07T18:39:50
Jupyter Notebook
UTF-8
R
false
false
495
r
helloworld.r
print("You can write any R code and package it into a container") print("Including your favorite packages") print("It will then run on the cloud") print("You can submit as many jobs as you like") print("And tear down the infrastructure after its finished") print("So that you can safe some money :)") print(sum(5,5)) print(sum(5,4)) print(sum(5,3)) print(sum(5,2)) print(sum(5,1)) print(sum(3,2)) print(sum(3,1)) print(sum(2,1)) print(sum(1,1)) print(sum(1,0)) print(sum(0,0)) print("Bye!!!!")
6685234d936cab551283ea7d12e5a670777591cf
ffdea92d4315e4363dd4ae673a1a6adf82a761b5
/data/genthat_extracted_code/sigora/examples/reaM.Rd.R
28ee55278c0e8e7f3fbe9d81b45f5746f6e81ead
[]
no_license
surayaaramli/typeRrh
d257ac8905c49123f4ccd4e377ee3dfc84d1636c
66e6996f31961bc8b9aafe1a6a6098327b66bf71
refs/heads/master
2023-05-05T04:05:31.617869
2019-04-25T22:10:06
2019-04-25T22:10:06
null
0
0
null
null
null
null
UTF-8
R
false
false
190
r
reaM.Rd.R
library(sigora) ### Name: reaM ### Title: Pathway GPS data, extracted from Reactome repository (Mouse). ### Aliases: reaM ### Keywords: datasets ### ** Examples data(reaM) str(reaM)
d8ae0d55fde5f921ae412109224fc95f35e44d29
c9dcad8a10c0e8f7571dda9460fa535e105919b9
/Experiments/Tibshirani2013/SGL/R/predict.SGL.R
c02faa4f93f1d92db60eeedbb125a14119166889
[]
no_license
adityagc/MS-Research
f0fd57420768ac270dd6007fa3ed015b5878e17c
5f148103a6205092fb3ae114c8613f7b6e849a84
refs/heads/master
2020-03-18T16:34:47.512135
2018-07-02T19:24:50
2018-07-02T19:24:50
134,974,224
3
0
null
null
null
null
UTF-8
R
false
false
493
r
predict.SGL.R
predict.SGL = function(x,newX,lam){ cvobj = x X <- newX if(!is.null(x$X.transform)){ X <- t(t(newX) - x$X.transform$X.means) X <- t(t(X) / x$X.transform$X.scale) } intercept <- 0 if(!is.null(x$intercept)){ intercept <- x$intercept[lam] } eta <- X %*% x$beta[,lam] + intercept if(x$type == "linear"){ y.pred <- eta } if(x$type == "logit"){ y.pred = exp(eta)/(1+exp(eta)) } if(x$type == "cox"){ y.pred = exp(eta) } return(y.pred) }
e6caf8b83ae33dd9657e3a27391da1d4ffda5740
da584e949c39b879460632fd42c8a3916248bc91
/Clean_code/Clean_Code7.R
23440ca558ac9348cde751bd8437feb8981ac9a3
[]
no_license
JordanEvan77/Population-Spread
d57c16ba99acee13e6079b50cb12ce0d901796eb
26781e09434a411bccd84b12100cb5d7781854f5
refs/heads/main
2023-04-27T09:42:47.310415
2021-05-14T19:13:16
2021-05-14T19:13:16
367,454,708
0
0
null
null
null
null
UTF-8
R
false
false
2,511
r
Clean_Code7.R
library(tidyverse) library(haven) #loading libraries for data manipulation #Question 1 and 2 Dimensions of data nfhs <- read_dta('Raw_data/IAHR52FL.dta') #This is the full initial raw data #Question 3, variables between "hhid" and "shstruc". nfhs_reduced <- select(nfhs, hhid:shstruc) %>% rename(survey_month = hv006) %>% rename(loc_type = hv026) #much easier to view table, less columns, only necessary data. nfhs_urban <- select(nfhs, hhid:shstruc) %>% rename(survey_month = hv006) %>% rename(home_loc = hv025) %>% rename(loc_type = hv026) %>% filter(home_loc == 1)#smaller, urban only, rural dropped #Question 4, Plot the distribution of the number of listed household members #for the entire sample. ggplot(data = nfhs_reduced, mapping = aes(x = hv009), binwidth = 1) + geom_histogram() + xlab("Number of household members") #Works! simple bar plot, skewed to right showing full data distribution. #QUestion 5, Create a boxplot plot using the data frame for urban area. #FACTOR: nfhs_1 <-as.factor(nfhs_urban$loc_type) #factor type for sorting urban only #PLOT: urban_plot <- ggplot(nfhs_urban) + aes(x = nfhs_1, y = hv009) urb_labels <- c("Large City", "Small city", "Town", "Country Side", "Missing") urban_plot + geom_boxplot() + xlab("Home Location") + ylab("Number Of Household Members") + scale_x_discrete(labels = urb_labels)#very nice box plot for showing the house member counts per location type. #Question 6,Use "group_by" and "summarise" to find the means and medians of the number of household members #by type of urban area. nfhs_urban %>% group_by(loc_type) %>% summarise_at(vars(hv009), list(name=mean))#list of means should show below nfhs_urban %>% group_by(loc_type) %>% summarise_at(vars(hv009), list(name=median))#list of medians should show as below #LIST OF MEANS PEOPLE PER HOUSE #[capital, large city] 4.65 #[small city] 4.88 #[town] 4.69 #LIST OF MEDIAN PEOPLE PER HOUSE #[capital, large city] 4 #[small city] 4 #[town] 4 #Question 7, What does the relationship between the mean and median tell you about #the distribution of household size? #The distribution doesn't vary much. This might indicate that large intergenerational homes #aren't the norm for the urban families, because the means are all within .23 "people" of eachother #and the median is the same for each location type. Additional comments on quiz.
f7e5b9a64adef987d4f06f4f5c1a1059a334e447
1486b7de41ac7a9882ecd92722d886802d49f9aa
/create_data/oceans/create_oceans.R
6c97a3f5ee78765f71af6e491f77c1442ab98304
[]
no_license
richardsc/ocedata
4f155f7ad72ac90a42c6003fb316c5e41b7e5a8b
ad804a151f3d50ea41df6f7cfd401e0446ddac6a
refs/heads/master
2021-01-01T17:58:48.937286
2017-07-24T16:20:25
2017-07-24T16:20:25
98,209,705
0
0
null
2017-07-24T16:08:07
2017-07-24T16:08:07
null
UTF-8
R
false
false
221
r
create_oceans.R
# http://ngdc.noaa.gov/mgg/global/etopo1_ocean_volumes.html oceans <- read.table("oceans.dat") # make area m^2 oceans$Area <- oceans$Area * 1e6 save(oceans, file="oceans.rda") tools::resaveRdaFiles(".", compress="auto")
e2e623032d56ca56c336418e7c33037cc7911449
caca7203b28507ec914b0be0042f96eb66db71ab
/code/study_var_significativity.R
2d37cb5f02a4abfa06947ee4105104f2f0a7e405
[]
no_license
boulbi777/returns-to-schooling-on-earning
e74e86ef94e8e7439a329f5239648746c8b67c6e
b9a64399e0cdf6ccc743276e30c99a59d81d58d0
refs/heads/master
2022-12-19T21:53:07.746490
2020-09-04T14:55:51
2020-09-04T14:55:51
293,286,960
0
0
null
null
null
null
UTF-8
R
false
false
494
r
study_var_significativity.R
#significativity test : 15-20% print("age :") var_significativity(data$age) #ok print("age2 :") var_significativity(data$age2) #ok print("female :") var_significativity(data$female) #ok print("for. :") var_significativity(data$for.) #ok print("reg1 :") var_significativity(data$reg1, FALSE) #ok print("reg2 :") var_significativity(data$reg2, FALSE) #not ok print("reg3 :") var_significativity(data$reg3, FALSE) #not really ok print("reg4 :") var_significativity(data$reg4, FALSE) #ok
47b5e7d25cd2267cfafb422f0a1acb832f5518b6
360bf7d821f1338c3d1e489c99608bfaaff97d26
/man/aggregate_points_space_time.Rd
4acbd01174ea749d602e0fbe0fed3fae0cde9816
[ "MIT" ]
permissive
disarm-platform/disarmr
0cd17c7e0270b2baa6036cf126d418fc9a39d9d2
a43b01d77d3552f8353caf9ae542a6798a25a384
refs/heads/master
2020-08-26T22:30:07.855368
2020-02-27T15:56:09
2020-02-27T15:56:09
217,168,297
1
1
MIT
2020-02-27T16:01:18
2019-10-23T22:59:12
R
UTF-8
R
false
true
377
rd
aggregate_points_space_time.Rd
% Generated by roxygen2: do not edit by hand % Please edit documentation in R/space_time_ppmify_helpers.R \name{aggregate_points_space_time} \alias{aggregate_points_space_time} \title{The aggregate_points_space_time function} \usage{ aggregate_points_space_time(points, ppmx, periods, date_start_end, reference_raster) } \description{ Helper function for space_time_ppmify }
613d525934deba71f76dacd2552bc43d1ac17abc
fcaf0c32ae844cb182dcb32f61649e2e81493809
/man/transitive.closure.Rd
c8f12703bfc09041f04f7efe679bbf588d320b08
[]
no_license
cran/clusthaplo
2d855b36fa4961bd70ec8dc20b7f847f1e7aa8e5
7bb3a732d17178223f0ae6ce0bcf359af15c9765
refs/heads/master
2016-09-06T08:43:09.813511
2013-07-05T00:00:00
2013-07-05T00:00:00
null
0
0
null
null
null
null
UTF-8
R
false
false
1,310
rd
transitive.closure.Rd
\name{transitive.closure} \alias{transitive.closure} \title{Performs transitive closure.} \usage{ transitive.closure(pair.simi, Npar=ncol(.gen)) } \note{ This function can only be run from inside a clusthaplo context as returned by clusthaplo(...). } \arguments{ \item{pair.simi}{The result of a call to \code{\link{pairwise.similarities}}.} \item{Npar}{The number of haplotypes for which the pairwise similarities were computed.} } \description{ Performs transitive closure on the whole set of pairwise similarities for one chromosome. Transitive closure means that, at a given locus, if haplotypes A and B match, and A and C match, then B and C are assumed to match also. } \value{ A matrix with one column per parent and one row per scanned locus, giving, at each locus and for each parent, the smallest parent index belonging in their clique. For instance, if haplotypes 1 3 4 look alike amidst 5 haplotypes, and 2 and 5 are singletons, the row will be \code{c(1, 2, 1, 1, 5)}. } \seealso{ \code{\link{pairwise.similarities}} } \examples{ data(parents.map, parents.gen) clu <- clusthaplo(parents.map, NULL, parents.gen) clu$select.chromosome('chr1') clu$train() tc <- clu$transitive.closure(clu$pairwise.similarities()) print(head(tc)) }
2de68f7654b45b51edc77aa4c1a73a504a9c5222
1d7ed56a3d9b6616509e901dbdac6728e6102bf2
/R/MergeC.R
1d2957285ceb173adcbc3e42f5a8414493989376
[]
no_license
EricMarcon/entropart
082f32c6c35b964d39c46fbd6b3c49f1b207d1fe
a24fd0cf8cd19896bf2db21e5a619dcdf22c1ae4
refs/heads/master
2023-07-20T14:16:47.132287
2023-07-14T09:01:21
2023-07-14T09:01:21
95,141,768
6
5
null
2020-11-09T19:23:35
2017-06-22T17:49:13
R
UTF-8
R
false
false
1,002
r
MergeC.R
MergeC <- function(MClist, Weights = rep(1, length(MClist)), CheckArguments = TRUE) { if (CheckArguments) CheckentropartArguments() # Metacommunities must have names if (is.null(names(MClist))) names(MClist) <- paste("MC", seq_along(MClist), sep="") CommunityNames <- function(MClist) { MCnames <- rep(names(MClist), unlist(lapply(MClist, function(x) length(x$Ni)))) paste(MCnames, unlist(lapply(MClist, function(x) names(x$Ni))), sep=".") } # Merge metacommunities Nsi Reduce(function(...) mergeandlabel(...), lapply(MClist, function(x) x$Nsi)) -> Gabundances NumCommunities <- unlist(lapply(MClist, function(x) length(x$Ni))) MCnames <- rep(names(MClist), NumCommunities) names(Gabundances) <- paste(MCnames, unlist(lapply(MClist, function(x) names(x$Ni))), sep=".") MCWeights <- unlist(lapply(MClist, function(x) x$Wi)) # Create the global MC return(MetaCommunity(Gabundances, MCWeights*rep(Weights, NumCommunities))) }
a98aa35b5f01ab4a41d5acf63e532a39d7041b97
323e8daf6ec798daf4527280e65473f10ae6ba16
/transform.R
1eb396cc0976b3ba709f103f924f4e163c98ffb9
[]
no_license
stjordanis/util-1
41812e236a93c2ddb1132228578bdbb41ac8d7bb
413433da0dbacca86bb0ad80b84c836a3aadf4bc
refs/heads/master
2021-12-02T20:10:14.899922
2012-11-05T20:17:15
2012-11-05T20:17:15
null
0
0
null
null
null
null
UTF-8
R
false
false
696
r
transform.R
##' Turn a factor in multiple dichotom factors ##' ##' This function takes a vector with an encoded factor and splits it up in a data.frame ##' with one column for each factor level. The value is one if the factor has the certain level, else ##' it is 0. ##' @title Dichotomize a factor ##' @param fac A factor vector ##' @return A data.frame with columns for each factor level ##' @author chris dichotomize <- function (fac, name){ fac_unique <- unique(fac) df <- as.data.frame(sapply(fac_unique, function(x) { ifelse(fac == x, 1, 0) })) colnames(df) <- paste(name, fac_unique, sep = "_") df }
2d53660fd26d77d39fb6d033f91a6b85dfe14d4e
7688e8e60019de7fca6fed79465904d765dcccd1
/Assignment_KNN/KNN-zoo.R
02a1368e08227fbbc88584e3162f865cfbff07e3
[]
no_license
AmitkumarShende/Assignments
35e12bc64bd06781695c3f8e6e1f58eda522d426
5402d022cda0343831f5f132d29f04c2f3bbe9ea
refs/heads/master
2021-02-09T22:53:37.847942
2020-03-02T09:37:03
2020-03-02T09:37:03
244,331,870
0
0
null
null
null
null
UTF-8
R
false
false
1,284
r
KNN-zoo.R
#**************install the packages if unavailable************** #install.packages('caret') #install.packages('dplyr') library(caret) # Read the dataset zoo <- read.csv(file.choose()) #EDA table(zoo$type) summary(zoo) str(zoo) # excluding 1st column having categorical values zoo1 <- zoo[,2:18] str(zoo1) #converting int variable type to factor type library(dplyr) con.names = zoo1 %>% select_if(is.numeric) %>% colnames() #con.names zoo1[,con.names] = data.frame(apply(zoo1[con.names], 2, as.factor)) str(zoo1) # Data partition set.seed(123) ind <- sample(2,nrow(zoo1), replace = T, prob = c(0.7,0.3)) train <- zoo1[ind==1,] test <- zoo1[ind==2,] #Creating performance Model # KNN Model trcontrol <- trainControl(method = "repeatedcv", number = 10,repeats = 3 # classprobs are needed when u want to select ROC for optimal K Value ) set.seed(123) fit <- train(type ~., data = train, method = 'knn', tuneLength = 20, trControl = trcontrol, preProc = c("center","scale")) # Default metric is accuracy but if u want to use ROC, then mention the same # Model Performance : fit # the optimum value for k should be 5 plot(fit) varImp(fit) pred <- predict(fit, newdata = test ) confusionMatrix(pred, test$type) #Accuracy of modelis 86.21
87ecbdeb9575196e886b8b28c792df4c59927753
5c0b9f40f6b93356c701941609850d90e2a56d15
/tests/testthat/test-make_wallpapr.R
8be7741330079d0c8aff366f1d7cf49f055147f9
[]
no_license
JBGruber/wallpapr
27710f46d3f3d76fd007c6f753effebe007d59e2
8ca68c982dbeccc40bff0f0ba5b90cbf0161f2a5
refs/heads/master
2022-10-12T18:31:09.090297
2022-10-04T09:43:29
2022-10-04T09:43:29
181,216,998
0
0
null
null
null
null
UTF-8
R
false
false
425
r
test-make_wallpapr.R
test_that("make_wallpapr", { expect_equal({ out <- make_wallpapr( system.file("extdata", "mull.jpg", package = "wallpapr"), month = as.Date("2021-03-01"), return_plot = TRUE ) df <- out$data c(nrow(df), ncol(df), sum(df$week), sum(df$size), class(out), attr(out, "dims")) }, list(39L, 6L, 387, 40, "wallpapr", "gg", "ggplot", dpi = 900, width = 1613L, height = 907L)) })
6aeda57a86214362ef98d77066db8d88c5acccb2
bea5690609b4cc2aded036e104685af3d996abd3
/createsimulationsobject.R
798773b958d9bbb20bc59afb026a8d1b99631f1e
[]
no_license
dluzzi/spup-app
6cc7f336d8690c26c162932777d516397e2b5666
7355e147eea0ab6fd1a1df763118847a3b354ba9
refs/heads/master
2016-08-11T10:51:16.429093
2016-03-15T13:56:16
2016-03-15T13:56:16
47,685,043
0
0
null
null
null
null
UTF-8
R
false
false
1,450
r
createsimulationsobject.R
## Used for testing only ## Creates an obejct of class "simulation" library(raster) # Create Class ------------------------------------------------------------ # Create "Simulations" class setClass("Simulations", slots = c(Realisations="RasterBrick", Mean = "RasterLayer", Standard.Deviation = "RasterLayer", Most.Likely.Class = "RasterLayer", Class.Probabilities = "RasterBrick", Quantiles = "RasterBrick") ) # Load data --------------------------------------------------------------- # Initialise rasterbrick zlatibor.brick <- brick() # Load data to create object for (i in 1:100){ #input <- Insert directory #input <- paste("D:/DamianoLuzzi-Thesis-DO-NOT-REMOVE/spup/data/zlatibor_dem_simulations/DEMsim", i, ".asc", sep = "") DEM <- raster(input) zlatibor.brick <- addLayer(zlatibor.brick, DEM) } # Convert stack to brick zlatibor.brick <- brick(zlatibor.brick) # Calculate mean, sd and quantiles std<-calc(zlatibor.brick, fun = sd, na.rm = T) mean <- mean(zlatibor.brick, na.rm = T) quantiles <- calc(zlatibor.brick, fun = function(x) {quantile(x, probs = c(.05,.25, .5, .75, .95),na.rm=TRUE)} ) # Create object of class Simulations -------------------------------------- simulations <- new("Simulations", Realisations = zlatibor.brick, Mean = mean, Standard.Deviation = std, Quantiles = quantiles)
2edb076da127ecd6e6ba156becb0c9ac8cb58c21
f8dba00a0e35196f01a174002350a0dbee36cdbb
/man/UKBBcleanR-package.Rd
f565abeefdda584290332f4e4a38130c27841e61
[ "MIT" ]
permissive
machiela-lab/UKBBcleanR
9801079ab95a88d0508113fb4d11544475b7a699
73c76b46258d810d936bdc5f4a3ec8b7d51cea41
refs/heads/main
2023-04-07T03:38:31.741924
2023-01-26T14:53:15
2023-01-26T14:53:15
520,682,551
5
0
null
null
null
null
UTF-8
R
false
true
1,425
rd
UKBBcleanR-package.Rd
% Generated by roxygen2: do not edit by hand % Please edit documentation in R/package.R \docType{package} \name{UKBBcleanR-package} \alias{UKBBcleanR-package} \alias{UKBBcleanR} \title{The UKBBcleanR Package} \description{ Prepare electronic medical record data from the UK Biobank for time-to-event analyses } \details{ Prepares time-to-event data from raw UK Biobank \url{https://www.ukbiobank.ac.uk/} electronic medical record data. The prepared data can be used for cancer outcomes but could be modified for other health outcomes. Key content of the 'UKBBcleanR' package include:\cr \code{\link{tte}} Prepares time-to-event data from raw UK Biobank \url{https://www.ukbiobank.ac.uk/} electronic medical record data. } \section{Dependencies}{ The 'UKBBcleanR' package relies heavily upon \code{\link{data.table}}, \code{\link{dplyr}}, and \code{\link{stringr}} to clean raw UK Biobank data \url{https://www.ukbiobank.ac.uk/} and output a time-to-event data set. } \author{ Alexander Depaulis\cr \emph{Integrative Tumor Epidemiology Branch (ITEB), Division of Cancer Epidemiology and Genetics (DCEG), National Cancer Institute (NCI), National Institutes of Health (NIH), Rockville, Maryland (MD), USA} \cr Derek W. Brown\cr \emph{ITEB, DCEG, NCI, NIH, Rockville, MD, USA} \cr Aubrey K. Hubbard\cr \emph{ITEB, DCEG, NCI, NIH, Rockville, MD, USA} \cr Maintainer: D.W.B. \email{derek.brown@nih.gov} } \keyword{package}
6f6d6e194218735dca1af2807a28401ac5bd778b
44f56ff45d793bc38f638f2fc88afd21893c72ba
/man/get_fpr.Rd
dcc1d932df3a9453faec0e8985f1d8118376d0d9
[]
no_license
pedrostrusso/annotator
e5fbeef5f954f4abd78a66985b740676d96c6c9f
4370b0e23395362db12a671e1574dd8bc525c0ac
refs/heads/master
2020-03-18T15:06:52.590658
2019-06-10T22:00:14
2019-06-10T22:00:14
134,887,810
0
0
null
null
null
null
UTF-8
R
false
true
547
rd
get_fpr.Rd
% Generated by roxygen2: do not edit by hand % Please edit documentation in R/tpr_fpr.R \name{get_fpr} \alias{get_fpr} \title{Calculate false positive rate between two clustering methods} \usage{ get_fpr(group1, group2) } \arguments{ \item{group1}{The first clustering method} \item{group2}{The reference ("true") method} } \value{ The false positive rate } \description{ Calculate false positive rate between two clustering methods } \examples{ g1 <- sample(1:2, size=10, replace=TRUE) g2 <- sample(1:3, size=10, replace=TRUE) get_fpr(g1, g2) }
0b092099c5873dfb3f0123a983782436b14e239d
e1967b6cf66fd951e75f10d9e2b4e68da6efce2c
/quickstart.R
52f9dade73dbaf264b19fbc903e72ab0047edd03
[]
no_license
PhilippPro/surrogate
f1e45e30b4e0ca2330b9a73f61cc8d274fad2fec
0e9d0b0252f6e200efa0208ad7ad3dbc167454cf
refs/heads/master
2021-01-19T00:22:43.953683
2017-04-06T13:43:23
2017-04-06T13:43:23
87,162,451
0
0
null
null
null
null
UTF-8
R
false
false
1,702
r
quickstart.R
library(devtools) # replace this soon load_all("/home/probst/Paper/Exploration_of_Hyperparameters/OMLbots") # This has to be replaced by the database extraction (Daniel) ---------------------------------------------- tag = "mlrRandomBot" numRuns = 140000 results = do.call("rbind", lapply(0:floor(numRuns/10000), function(i) { return(listOMLRuns(tag = tag, limit = 10000, offset = (10000 * i) + 1)) }) ) table(results$flow.id, results$task.id) table(results$uploader) res = do.call("rbind", lapply(0:floor(nrow(results)/100), function(i) { return(listOMLRunEvaluations(run.id = results$run.id[((100*i)+1):(100*(i+1))])) }) ) # dauert ewig df = res %>% mutate(flow.version = c(stri_match_last(flow.name, regex = "[[:digit:]]+\\.*[[:digit:]]*")), learner.name = stri_replace_last(flow.name, replacement = "", regex = "[([:digit:]]+\\.*[[:digit:]*)]")) as.data.frame.matrix(table(df$learner.name, df$data.name)) # ----------------------------------------------------------------------------------------------------------- overview = getMlrRandomBotOverview("botV1") print(overview) tbl.results = getMlrRandomBotResults("botV1") print(tbl.results) tbl.hypPars = getMlrRandomBotHyperpars("botV1") print(tbl.hypPars) task.data = makeBotTable(measure.name = "area.under.roc.curve", learner.name = "mlr.classif.glmnet", tbl.results = tbl.results, tbl.hypPars = tbl.hypPars, tbl.metaFeatures = NULL) task.ids = unique(tbl.results$task.id) surr = makeSurrogateModel(measure.name = "area.under.roc.curve", learner.name = "mlr.classif.glmnet", task.id = task.ids, tbl.results = tbl.results, tbl.hypPars = tbl.hypPars, param.set = lrn.par.set$classif.glmnet.set$param.set)
51fa4fb4dc766f9cd2d2033fe9fb1e345eb2fcdf
b62a6e662910152cf4e41e8e98c3b090f69ae245
/pould/R/parseGenotypes.R
242715ea9e5764c113dd68fee461502130c37a81
[]
no_license
sjmack/pould
b359d3f6cf44fc858dfd6666a9e2771b7d09bd09
bd35da7715d6faa960620f3bf6816031d8c8d9b3
refs/heads/master
2021-06-29T20:53:53.423809
2020-10-08T18:30:00
2020-10-08T18:30:00
132,675,243
1
2
null
2018-05-19T07:58:12
2018-05-08T23:15:13
R
UTF-8
R
false
false
4,357
r
parseGenotypes.R
## parseGenotypes -- Steven J. Mack April 10, 2020 ## v1.00 ## Accepts and converts 2-column/locus BIGDAWG/PyPop-formatted genotype data to the GL String format expected by LDWrap #' Reformat columnnar genotype data to GL String format #' #' This function accepts genotype data organized in locus-column pairs, and returns GL String-formatted data structured for LDWrap(). Of the resulting multilocus haplotype pair, the first haplotype is constructed from the first column for each locus, and the second haplotype is constructed from the second column. #' @param dataset A tab-delimited text file (with a .txt or .tsv filename suffix) with a header row or a data frame. Each row corresponds to a subject, with two columns per locus. Allele names can include a locus name (e.g., locus*allele) or can can exclude the locus, but all allele names in the dataset must either include or exclude the locus. Missing (untyped) allele data can be identified with an empty cell or a set of four asterisks in files, and with NA values in data frames. Column names for each locus pair must be adjacent, but can be either identical (e.g., "locus" and "locus"), or suffixed (e.g., "locus_1" and "locus_2", where "locus_1" always precedes "locus_2"). A optional column of sample identifiers can be included, but must be named "SampleID". A column named "Disease" can be included, but will be ignored. No other non-locus columns are permitted. #' @note This function is for internal POULD use only. #' @return A data frame of two columns. The "Relation" column includes sample identifiers if provided, or numbers from 1 to the number of subjects. The "GL String" column contains the GL String formatted genotypes. #' @keywords LDformat reformat GL String #' @export #' @examples # parseGenotypes <- function(dataset) { if(missing(dataset)) {return(cat("Please provide a value for the dataset parameter.\n"))} if(!is.data.frame(dataset)) { dataset <- read.table(dataset,header=T,sep="\t",colClasses = "character",stringsAsFactors = FALSE,as.is = TRUE,check.names = FALSE,na.strings = "****")} colnames(dataset) <- toupper(colnames(dataset)) if("SAMPLEID" %in% colnames(dataset)) { ids <- dataset$SAMPLEID dataset <- dataset[,!colnames(dataset) %in% "SAMPLEID"] } else { ids <- 1:nrow(dataset) } if("DISEASE" %in% colnames(dataset)) { dataset <- dataset[,!colnames(dataset) %in% "DISEASE"] } if(ncol(dataset) %% 2 !=0 ) {return(cat("Odd number of locus columns (",ncol(dataset),"). Please review your dataset.\n",sep=""))} colnames(dataset) <- sub("\\_\\d","",colnames(dataset)) if(ncol(dataset) == 2) {return(cat("This dataset contains data for a single locus (",colnames(dataset)[1],"). LD analysis requires two loci.\n",sep=""))} if(!any(grepl("*",dataset,fixed=TRUE))) {dataset[] <- Map(paste,names(dataset),dataset,sep="*")} # V0.3 remove NAs that become locus*NA blanks <- paste(colnames(dataset),NA,sep="*") for(i in 1:ncol(dataset)) { if(nrow(dataset[dataset[,i] == blanks[i],][i]) != 0 ) { dataset[dataset[,i] == blanks[i],][i] <- NA } } hap <-vector("list",2) # paste together haplotypes & clean up stragglers for(x in FALSE:TRUE) { hap[[((1*x)+1)]] <- apply(dataset[,rep(c(TRUE,FALSE),(ncol(dataset)/2))==x],1,paste,collapse="~") hap[[((1*x)+1)]] <- gsub("[N][A]","",hap[[((1*x)+1)]]) # eliminate all 'NA' from missing data cells hap[[((1*x)+1)]] <- gsub("~+","~",hap[[((1*x)+1)]]) # eliminate tilde-runs for empty cells hap[[((1*x)+1)]][substr(hap[[((1*x)+1)]],1,1)=="~"] <- substr(hap[[((1*x)+1)]][substr(hap[[((1*x)+1)]],1,1)=="~"],2,nchar(hap[[((1*x)+1)]][substr(hap[[((1*x)+1)]],1,1)=="~"])) ## trim leading tilde hap[[((1*x)+1)]][substr(hap[[((1*x)+1)]],(nchar(hap[[((1*x)+1)]])),nchar(hap[[((1*x)+1)]]))=="~"] <- substr(hap[[((1*x)+1)]][substr(hap[[((1*x)+1)]],(nchar(hap[[((1*x)+1)]])),nchar(hap[[((1*x)+1)]]))=="~"],1,nchar(hap[[((1*x)+1)]])-1) ## trim trailing tilde } fdataset <- cbind(as.data.frame(ids,stringsAsFactors = FALSE),as.data.frame(paste(hap[[2]],hap[[1]],sep="+"),stringsAsFactors = FALSE)) colnames(fdataset) <- c("Relation","Gl.String") fdataset }
e566fc8b712b52cb5cec57c3f5f487b6002db897
9eac9f8e7495d916f7596c4444461521b1a39086
/scripts/regionAnalysis.R
eac666de0313a555b3df26d5b091072296033db8
[ "Apache-2.0" ]
permissive
uniqueg/scripts
bbb42d455196f8e047df2681661a02d38e4a762f
9fdcb93f740c0d353b8f9c0fe3ceab6a941af87d
refs/heads/master
2023-04-08T17:08:00.911197
2023-03-16T08:46:40
2023-03-16T08:46:40
211,389,152
0
0
null
null
null
null
UTF-8
R
false
false
11,633
r
regionAnalysis.R
#!/usr/bin/env Rscript ############################## ### GLOBAL PARAMETERS // ### ############################## region_types <- c("three_prime_utr", "five_prime_utr", "CDS", "none") def_region_types <- c("three_prime_utr", "five_prime_utr", "CDS") reg_colors_def <- c("#33ccff", "#666666", "#ff8000") ############################## ### // GLOBAL PARAMETERS ### ############################## ###################### ### FUNCTIONS // ### ###################### formatTypes <- function(chr) { chr <- gsub("five_prime_utr", "5'UTR", chr) chr <- gsub("three_prime_utr", "3'UTR", chr) chr <- gsub("none", "Undefined", chr) return(chr) } #-----------------------# pieChart <- function(x, col, legend, main=NULL, cex.main=1.6, legend.pos="topright", cex.leg=1.4) { pie(x, clockwise=TRUE, col=col, labels=NA, main=main, cex.main=cex.main) legend(x=legend.pos, legend=legend, fill=col, bty="n", cex=cex.leg) return(NULL) } #-----------------------# chiSquare <- function(x, p) { # chsq <- chisq.test(x=x, p=p) meth <- chsq$method sep <- paste(rep("-", nchar(meth)), collapse="") stat <- paste("Statistic:", chsq$statistic) df <- paste("Degrees of freedom:", chsq$parameter) p <- paste("P-value:", chsq$p.value) obs <- paste("Observed:", paste(chsq$observed, collapse=", ")) exp <- paste("Expected:", paste(chsq$expected, collapse=", ")) res <- paste("Pearson residuals:", paste(chsq$residuals, collapse=", ")) stdres <- paste("Standardized residuals:", paste(chsq$stdres, collapse=", ")) chsq$print <- paste(meth, sep, stat, df, p, obs, exp, res, stdres, sep="\n") return(chsq) } #-----------------------# categoryVectorListToBED <- function(vectorList, nms, categories, start, prefix) { # grl <- GRangesList(mapply(function(name, start) { site <- vectorList[[name]] site_coll <- rle(site) type <- site_coll$values %in% categories site_coll <- setNames(site_coll$lengths[type], site_coll$values[type]) start <- start + c(0, cumsum(site_coll[-length(site_coll)])) ranges <- IRanges(start=start, width=site_coll) seqnames <- Rle(rep(name, length(ranges))) gr <- GRanges(seqnames=seqnames, ranges=ranges, name=names(site_coll)) }, nms, start)) gr <- unlist(grl) names(gr) <- NULL outFile <- paste(prefix, "regions.bed", sep=".") write(paste("Writing region types to BED file ", outFile, "...", sep="'"), stdout()) export(gr, outFile, format="bed") return(gr) } #-----------------------# plotFormats <- function(FUN, formats, prefix) { if ("pdf" %in% formats) { outFile <- paste(prefix, "pdf", sep=".") write(paste("Plotting pie chart to file ", outFile, "...", sep="'")) pdf(outFile) dump <- FUN dev.off() } # if ("png" %in% formats) { # outFile <- paste(prefix, "png", sep=".") # write(paste("Plotting pie chart to file ", outFile, "...", sep="'")) # png(outFile) # dump <- FUN # dev.off() # } # if ("svg" %in% formats) { # outFile <- paste(prefix, "svg", sep=".") # write(paste("Plotting pie chart to file ", outFile, "...", sep="'")) # svg(outFile) # dump <- FUN # dev.off() # } } #-----------------------# processAnnotations <- function(gtf, outDir) { # Load packages write("Loading package 'rtracklayer'...", stdout()) library("rtracklayer") # Make output directory & prepare output filenames write("Generating output directory...", stdout()) dir.create(outDir, recursive=TRUE, showWarnings = FALSE) gtfBase <- unlist(strsplit(basename(gtf), ".gtf"))[1] out_prefix <- file.path(outDir, gtfBase) # Import GTF annotations write(paste("Importing GTF annotation data from file ", gtf, "...", sep="'"), stdout()) gtf <- import(gtf, format="gtf", asRangedData=FALSE) outFile <- paste(out_prefix, "GRanges.R", sep=".") write(paste("Saving GTF object in R file ", outFile, "...", sep="'"), stdout()) save(gtf, file=outFile) # Subset exons and split by transcript identifier write("Subsetting exons...", stdout()) exons_gr <- gtf[gtf$type == "exon"] mcols(exons_gr) <- list(transcript_id=factor(exons_gr$transcript_id)) exons_grl <- split(exons_gr, exons_gr$transcript_id) # Subset & process region annotations (5' UTR, CDS, 3' UTR) write("Subsetting/processing region annotations...", stdout()) regions_gr <- gtf[gtf$type %in% c("five_prime_utr", "CDS", "stop_codon", "three_prime_utr")] regions_gr$type[regions_gr$type == "stop_codon"] <- "CDS" mcols(regions_gr) <- list(transcript_id=factor(regions_gr$transcript_id), type=factor(regions_gr$type)) regions_grl <- split(regions_gr, regions_gr$transcript_id) # Get exons with and without region annotation write("Identify exons with/without region annotations...", stdout()) exons_w_regions_grl <- exons_grl[intersect(names(exons_grl), names(regions_grl))] exons_wo_regions_grl <- exons_grl[setdiff(names(exons_grl), names(exons_w_regions_grl))] exons_w_partial_regions_grl <- psetdiff(exons_w_regions_grl, regions_grl) # Update region annotations with type "none" write("Add annotation type 'none' to unavailable/undefined region annotations...", stdout()) no_regions_gr <- c(unlist(exons_wo_regions_grl, use.names=FALSE), unlist(exons_w_partial_regions_grl, use.names=FALSE)) mcols(no_regions_gr)$type <- factor(rep("none", length(no_regions_gr))) regions_gr <- sort(c(regions_gr, no_regions_gr)) regions_grl <- split(regions_gr, regions_gr$transcript_id) # Get list of vectors of region types # List contains one vector for each transcript # Each vector is composed of region types for each position write("Obtaining region type information per nucleotide (this may take long)...", stdout()) reg_vec_all_ls <- lapply(regions_grl, function(trx) { strand <- unique(strand(trx)) if ( length(strand) != 1 | ! strand %in% c("+", "-") ) { write("[WARNING] Strand information unclear.", stderr()) return(NULL) } vec <- as.character(unlist(mapply(rep, x=trx$type, each=width(trx)))) if ( strand == "+" ) return(vec) else return(rev(vec)) }) outFile <- paste(out_prefix, "regionByNucleotide.R", sep=".") write(paste("Saving nucleotide-level composition information in file ", outFile, "...", sep="'"), stdout()) save(reg_vec_all_ls, file=outFile) # Summarize region type nucleotide composition write("Counting nucleotides per region type...", stdout()) reg_cts_all <- table(unlist(reg_vec_all_ls)) reg_cts_all <- setNames(as.numeric(reg_cts_all), names(reg_cts_all)) outFile <- paste(out_prefix, "regionCounts.R", sep=".") write(paste("Saving counts in file ", outFile, "...", sep="'"), stdout()) save(reg_cts_all, file=outFile) # Generate BED file of regions names_found <- names(reg_vec_all_ls[! sapply(reg_vec_all_ls, is.null)]) start <- rep(1, length(names_found)) gr <- categoryVectorListToBED(reg_vec_sites_ls, names_found, region_types, start, out_prefix) # Generate pie chart reg_cts_all_def <- reg_cts_all[def_region_types] plotFormats(pieChart(x=reg_cts_all_def[def_region_types], col=reg_colors_def, legend=formatTypes(def_region_types), main="control"), formats=c("pdf", "png", "svg"), prefix=paste(out_prefix, "regionCounts.pie", sep=".")) # Return list of objects obj_ls <- list(gtf=gtf, regions_gr=regions_gr, reg_vec_all_ls=reg_vec_all_ls, reg_cts_all=reg_cts_all, gr=gr, reg_cts_all_def=reg_cts_all_def) return(obj_ls) } #-----------------------# processSample <- function (csv, regionPerNt, reg_cts_all, outDir) { # Make output directory & prepare output filenames write("Generating output directory...", stdout()) dir.create(outDir, recursive=TRUE, showWarnings = FALSE) csvBase <- unlist(strsplit(basename(csv), ".csv"))[1] out_prefix <- file.path(outDir, csvBase) # Loading annotation data write(paste("Obtaining annotation R objects...", sep="'"), stdout()) if ( mode(regionPerNt) == "character" ) { load(regionPerNt) } else { reg_vec_all_ls <- regionPerNt } if ( mode(regionCounts) == "character" ) { load(regionCounts) } else { reg_cts_all <- regionCounts } # Importing CSV file of sites write(paste("Importing sites from file ", csv, "...", sep="'"), stdout()) sites <- read.delim(csv, stringsAsFactors=FALSE) # Get list of vectors of region types # List contains one vector for each transcript # Each vector is composed of region types for each position reg_vec_sites_ls <- apply(sites, 1, function(site) { reg_vec_all_ls[[site["seqnames"]]][site["start"]:site["end"]] }) names(reg_vec_sites_ls) <- sites$seqnames outFile <- paste(out_prefix, "regionByNucleotide.R", sep=".") write(paste("Saving nucleotide-level composition information in file ", outFile, "...", sep="'"), stdout()) save(reg_vec_sites_ls, file=outFile) # Summarize region type nucleotide composition write("Counting nucleotides per region type...", stdout()) reg_cts_sites <- table(unlist(reg_vec_sites_ls)) reg_cts_sites <- setNames(as.numeric(reg_cts_sites), names(reg_cts_sites)) outFile <- paste(out_prefix, "regionCounts.R", sep=".") write(paste("Saving counts in file ", outFile, "...", sep="'"), stdout()) save(reg_cts_sites, file=outFile) # Generate BED file of regions names_found <- names(reg_vec_sites_ls[! sapply(reg_vec_sites_ls, is.null)]) start <- sites$start[sites$seqnames %in% names_found] gr <- categoryVectorListToBED(reg_vec_sites_ls, names_found, region_types, start, out_prefix) # Generate pie chart reg_cts_sites_def <- reg_cts_sites[def_region_types] plotFormats(pieChart(x=reg_cts_sites_def, col=reg_colors_def, legend=formatTypes(def_region_types), main="sample"), formats=c("pdf", "png", "svg"), prefix=paste(out_prefix, "regionCounts.pie", sep=".")) # Run Pearson's Chi-squared test write("Running Pearson's Chi-squared test...", stdout()) reg_cts_all_def <- reg_cts_all[def_region_types] chsq <- chiSquare(x=reg_cts_sites_def, p=reg_cts_all_def/sum(reg_cts_all_def)) outFile <- paste(out_prefix, "chiSquare.txt", sep=".") write(paste("Writing Chi-square summary to file", outFile, "...", sep="'"), stdout()) write(chsq$print, file=outFile) # Return list of objects obj_ls <- list(reg_vec_all_ls=reg_vec_all_ls, reg_cts_all=reg_cts_all, sites=sites, reg_vec_sites_ls=reg_vec_sites_ls, reg_cts_sites=reg_cts_sites, reg_cts_sites_def=reg_cts_sites_def, reg_cts_all_def=reg_cts_all_def, gr=gr, chsq=chsq) return(obj_ls) } ###################### ### // FUNCTIONS ### ###################### ################# ### MAIN // ### ################# # Initiate objects obj_ls <- NULL # Process annotations if ( ! is.null(gtf) ) { annot_obj_ls <- processAnnotations(gtf, outDir) if ( is.null(regionPerNt) ) regionPerNt <- annot_obj_ls$reg_vec_all_ls if ( is.null(regionCounts) ) regionCounts <- annot_obj_ls$reg_cts_all } # Process sample if ( ! any(is.null(c(csv, regionPerNt, regionCounts))) ) { sample_obj_ls <- processSample(csv, regionPerNt, regionCounts, outDir) } # Save session outFile <- file.path(outDir, "session.R") write(paste("Saving R session in file ", outFile, "...", sep="'"), stdout()) save.image(file=outFile) ################# ### // MAIN ### #################
f3839bb2d1033962fb2109b54a723286636dbe23
7f9c3c40ec2ab8938e0482b2f07e2ceb1301b0be
/R/functions-simulation.R
c43047abdf5880303984426909224855623cb6ed
[]
no_license
graveja0/modeling-health-insurance
c7cfa157c4d15dfda76e39d5366e473042cf108a
ce1c10e205e0d5d026010b19b527af0d0749b89d
refs/heads/master
2021-07-10T09:04:41.948224
2020-07-21T14:31:58
2020-07-21T14:31:58
177,799,095
2
0
null
null
null
null
UTF-8
R
false
false
29,958
r
functions-simulation.R
theme_tufte_revised <- function(base_size = 11, base_family = "Gill Sans", ticks = TRUE) { ret <- ggplot2::theme_bw(base_family = base_family, base_size = base_size) + ggplot2::theme( axis.line = ggplot2::element_line(color = 'black'), axis.title.x = ggplot2::element_text(vjust = -0.3), axis.title.y = ggplot2::element_text(vjust = 0.8), legend.background = ggplot2::element_blank(), legend.key = ggplot2::element_blank(), legend.title = ggplot2::element_text(face="plain"), panel.background = ggplot2::element_blank(), panel.border = ggplot2::element_blank(), panel.grid = ggplot2::element_blank(), plot.background = ggplot2::element_blank(), strip.background = ggplot2::element_blank() ) if (!ticks) { ret <- ret + ggplot2::theme(axis.ticks = ggplot2::element_blank()) } ret } get_params <- function(x,params) { map2(params,x,~( .x(.y) )) } qfixed <- function(x, value) value get_takeup_coef <- function(df = df_wtp_and_costs,params) { sampled <- sample(df$iteration,1) tmp <- df %>% filter(iteration == sampled & fpl == params$pop_fpl & type == params$plan_type & outcome=="s") %>% select(-type,-outcome,-fpl,-iteration) tmp_coef <- tmp %>% gather(coef,value) %>% mutate(value = as.numeric(paste0(value))) %>% pull(value) names(tmp_coef) <- names(tmp) return(tmp_coef) } get_cost_coef <- function(df = df_wtp_and_costs, params) { sampled <- sample(df$iteration,1) tmp <- df %>% filter(iteration == sampled & fpl == params$pop_fpl & type == params$plan_type & outcome=="cost") %>% select(-type,-outcome,-fpl,-iteration) tmp_coef <- tmp %>% gather(coef,value) %>% mutate(value = as.numeric(paste0(value))) %>% pull(value) names(tmp_coef) <- names(tmp) return(tmp_coef) } get_takeup <- function(params, premium ) { if (is.null(premium)) prem = params$plan_premium else prem = premium tmp_out <- data.frame(fpl = params$pop_fpl, type = params$plan_type, prem = prem) %>% mutate(estimate = params$takeup_coef['intercept'] + params$takeup_coef['wtp'] * prem + params$takeup_coef['i_wtp_2'] * prem^2 + params$takeup_coef['i_wtp_3'] * prem ^3) %>% pull(estimate) %>% unname() return(pmax(0,pmin(1,tmp_out))) } get_cost <- function(params, premium ) { if (is.null(premium)) prem = params$plan_premium else prem = premium tmp_out <- data.frame(fpl = params$pop_fpl, type = params$plan_type, prem = prem) %>% mutate(estimate = params$cost_coef['intercept'] + params$cost_coef['wtp'] * prem + params$cost_coef['i_wtp_2'] * prem^2 + params$cost_coef['i_wtp_3'] * prem ^3) %>% pull(estimate) %>% unname() return(pmax(0,tmp_out)) } fn_uncomp <- function(cost, uninsured_oop_share , phi ) { # x is the share of the uninsured’s total health care costs that they pay out of pocket # φ denotes the percentage increase in costs that result from insurance coverage (moral hazard) (1 - uninsured_oop_share) * (cost / (1 + phi)) } TwoWaySA<-function(indata,outcome="NHB",parm1,parm2,range1,range2,lambda){ # Get Outcome lhs <- indata %>% select(psa_id,contains("dQALY"),contains("dCOST")) %>% mutate(psa_id=row_number()) %>% reshape2::melt(id.vars='psa_id') %>% tidyr::separate(variable,c("outcome","strategy"),"_") %>% reshape2::dcast(psa_id+strategy~outcome) %>% mutate(NHB = dQALY-dCOST * lambda , NMB = dQALY*lambda - dCOST) # Get Parameters rhs <- indata %>% select(-contains("dQALY"),-contains("dCOST"), -contains("NMB"),-contains("NHB"),-psa_id) # Map to existing code inputs Strategies <- unique(lhs$strategy) Parms <- rhs %>% tbl_df() %>% data.frame() cat(outcome) lhs$Y <- lhs[,outcome] Outcomes <- lhs %>% select(strategy,psa_id,Y) %>% reshape2::dcast(psa_id~strategy,value.var="Y") %>% select(-psa_id) #Extract parameter column number in Parms matrix x1<-which(colnames(Parms)==parm1) x2<-which(colnames(Parms)==parm2) dep<-length(Strategies) #Number of dependent variables, i.e., strategies indep<-ncol(Parms) #Number of independent variables, i.e., parameters Sim <- data.frame(Outcomes,Parms) if (ncol(Parms)==2) { Parms$constant = 1 Sim$constant = 1 } #Determine range of of the parameer to be plotted if (!missing("range1")&!missing("range2")){ #If user defines a range vector1<-seq(from=range1[1],to=range1[2],length.out=301) vector2<-seq(from=range2[1],to=range2[2],length.out=301) } else if (!missing("range1")&missing("range2")){ #Default range given by the domanin of the parameter's sample #vector to define 400 samples between the 2.5th and 97.5th percentiles vector1<-seq(from=range1[1],to=range1[2],length.out=301) y2 = seq(2.5,97.5,length.out=301) j2 = round(y2*(length(Parms[,x2])/100)) #indexing vector;j=round(y*n/100) where n is the size of vector of interest vector2<-sort(Parms[j2,x2]) } else if (missing("range1")&!missing("range2")){ #Default range given by the domanin of the parameter's sample #vector to define 400 samples between the 2.5th and 97.5th percentiles vector2<-seq(from=range2[1],to=range2[2],length.out=301) y1 = seq(2.5,97.5,length.out=301) j1 = round(y1*(length(Parms[,x1])/100)) #indexing vector;j=round(y*n/100) where n is the size of vector of interest vector1<-sort(Parms[j1,x1]) } else{ y1 = seq(2.5,97.5,length.out=301) y2 = seq(2.5,97.5,length.out=301) j1 = round(y1*(length(Parms[,x1])/100)) #indexing vector;j=round(y*n/100) where n is the size of vector of interest j2 = round(y2*(length(Parms[,x2])/100)) vector1<-sort(Parms[j1,x1]) vector2<-sort(Parms[j2,x2]) } #Generate a formula by pasting column names for both dependent and independent variables f <- as.formula(paste('cbind(',paste(colnames(Sim)[1:dep],collapse=','), ') ~ (','poly(',parm1,',8)+','poly(',parm2,',8)+' , paste(colnames(Parms)[c(-x1,-x2)], collapse='+'),')')) #Run Multiple Multivariate Regression (MMR) Metamodel Tway.mlm = lm(f,data=Sim) TWSA <- expand.grid(parm1=vector1,parm2=vector2) #Generate matrix to use for prediction Sim.fit<-matrix(rep(colMeans(Parms)),nrow=nrow(TWSA),ncol=ncol(Parms), byrow=T) Sim.fit[,x1]<-TWSA[,1] Sim.fit[,x2]<-TWSA[,2] Sim.fit<-data.frame(Sim.fit) #Transform to data frame, the format required for predict colnames(Sim.fit)<-colnames(Parms) #Name data frame's columns with parameters' names #Predict Outcomes using MMMR Metamodel fit Sim.TW = data.frame(predict(Tway.mlm, newdata = Sim.fit)) #Find optimal strategy in terms of maximum Outcome Optimal <- max.col(Sim.TW) #Get Outcome of Optimal strategy OptimalOut<-apply(Sim.TW,1,max) plotdata = Sim.fit #Append parameter's dataframe to predicted outcomes dataframe #A simple trick to define my variables in my functions environment plotdata$parm1<-plotdata[,parm1]; plotdata$parm2<-plotdata[,parm2]; plotdata$Strategy<-factor(Optimal,labels=Strategies[as.numeric(names(table(Optimal)))]) plotdata$value<-OptimalOut txtsize<-12 p <- ggplot(plotdata, aes(x=parm1,y=parm2))+ geom_tile(aes(fill=Strategy)) + theme_bw() + #ggtitle(expression(atop("Two-way sensitivity analysis", # atop("Net Health Benefit")))) + scale_fill_discrete("Strategy: ", l=50)+ xlab(parm1)+ ylab(parm2)+ theme(legend.position="bottom",legend.title=element_text(size = txtsize), legend.key = element_rect(colour = "black"), legend.text = element_text(size = txtsize), title = element_text(face="bold", size=15), axis.title.x = element_text(face="bold", size=txtsize), axis.title.y = element_text(face="bold", size=txtsize), axis.text.y = element_text(size=txtsize), axis.text.x = element_text(size=txtsize))+ scale_fill_grey(start = 0, end = 1) return(p) } OneWaySA<-function(indata,outcome="NHB",lambda,parm,range){ # Get Outcome lhs <- indata %>% select(psa_id,contains("dQALY"),contains("dCOST")) %>% mutate(psa_id = row_number()) %>% reshape2::melt(id.vars='psa_id') %>% tidyr::separate(variable,c("outcome","strategy"),"_") %>% reshape2::dcast(psa_id+strategy~outcome) %>% mutate(NHB = dQALY-dCOST * lambda , NMB = dQALY*lambda - dCOST) # Get Parameters rhs <- indata %>% select(-contains("dQALY"),-contains("dCOST"), -contains("NMB"),-contains("NHB"),-psa_id) # Map to existing code inputs Strategies <- unique(lhs$strategy) Parms <- rhs %>% tbl_df() %>% data.frame() lhs$Y <- lhs[,outcome] Outcomes <- lhs %>% select(strategy,psa_id,Y) %>% reshape2::dcast(psa_id~strategy,value.var="Y") %>% select(-psa_id) #Extract parameter column number in Parms matrix x<-which(colnames(Parms)==parm) dep<-length(Strategies) #Number of dependent variables, i.e., strategies outcomes indep<-ncol(Parms) #Number of independent variables, i.e., parameters Sim <- data.frame(Outcomes,Parms) #Determine range of of the parameer to be plotted if (!missing("range")){ #If user defines a range vector<-seq(range[1],range[2],length.out=400) } else{ #Default range given by the domanin of the parameter's sample #vector to define 400 samples between the 2.5th and 97.5th percentiles y = seq(2.5,97.5,length=400) j = round(y*(length(Parms[,x])/100)) #indexing vector;j=round(y*n/100) where n is the size of vector of interest vector<-sort(as.data.frame(Parms)[j,x]) } #Generate a formula by pasting column names for both dependent and independent variables. Imposes a 1 level interaction f <- as.formula(paste('cbind(',paste(colnames(Sim)[1:dep],collapse=','), ') ~ (','poly(',parm,',2)+' ,paste(colnames(Parms)[-x], collapse='+'),')')) #Run Multiple Multivariate Regression (MMR) Metamodel Oway.mlm = lm(f,data=Sim) #Generate matrix to use for prediction Sim.fit<-matrix(rep(colMeans(Parms)),nrow=length(vector),ncol=ncol(Parms), byrow=T) Sim.fit[,x]<-vector Sim.fit<-data.frame(Sim.fit) #Transform to data frame, the format required for predict colnames(Sim.fit)<-colnames(Parms) #Name data frame's columns with parameters' names #Predict Outcomes using MMMR Metamodel fit plotdata = data.frame(predict(Oway.mlm, newdata = Sim.fit)) colnames(plotdata) <- Strategies #Name the predicted outcomes columns with strategies names #Reshape dataframe for ggplot plotdata = stack(plotdata, select=Strategies) # plotdata = cbind(Sim.fit, plotdata) #Append parameter's dataframe to predicted outcomes dataframe #A simple trick to define my variables in my functions environment plotdata$parm<-plotdata[,parm]; library(directlabels) txtsize<-12 #Text size for the graphs ggplot(data = plotdata, aes(x = parm, y = values, lty = ind)) + geom_line() + #ggtitle("One-way sensitivity analysis \n Net Health Benefit") + xlab(parm) + ylab("E[NHB]") + scale_colour_hue("Strategy", l=50) + #scale_x_continuous(breaks=number_ticks(6)) + #Adjust for number of ticks in x axis #scale_y_continuous(breaks=number_ticks(6)) + theme_bw() + theme(legend.position="bottom",legend.title=element_text(size = txtsize), legend.key = element_rect(colour = "black"), legend.text = element_text(size = txtsize), title = element_text(face="bold", size=15), axis.title.x = element_text(face="bold", size=txtsize), axis.title.y = element_text(face="bold", size=txtsize), axis.text.y = element_text(size=txtsize), axis.text.x = element_text(size=txtsize))+ geom_dl(aes(label = ind), method = list(dl.combine("last.bumpup"), cex = 0.8)) } CEAC<-function(lambda_range,indata){ # Get Outcome lhs <- indata %>% select(contains("dQALY"),contains("dCOST")) %>% mutate(psa_id = row_number()) %>% reshape2::melt(id.vars='psa_id') %>% tidyr::separate(variable,c("outcome","strategy"),"_") %>% reshape2::dcast(psa_id+strategy~outcome) # Get Parameters rhs <- indata %>% select(-contains("dQALY"),-contains("dCOST"),-psa_id) # Map to existing code inputs Strategies <- unique(lhs$strategy) Parms <- rhs %>% tbl_df() Outcomes <- lhs %>% select(strategy,psa_id,contains("dCOST"),contains("dQALY")) # Outcomes must be ordered in a way that for each strategy the cost must appear first then the effectiveness lambda<- lambda_range NHB <- array(0, dim=c(dim(Outcomes)[1],length(Strategies))) # Matrix to store NHB for each strategy colnames(NHB)<-Strategies CEA<-array(0,dim=c(length(lambda),length(Strategies))) # NHB <- lambda %>% purrr::map(~(Outcomes$dQALY-Outcomes$dCOST * .x)) %>% do.call("cbind",.) colnames(NHB) <- paste0("lambda_",lambda) NHB <- data.frame(NHB) NHB$strategy <- Outcomes$strategy NHB$psa_id <- Outcomes$psa_id NHB2 <- NHB %>% reshape2::melt(id.vars=c("strategy","psa_id")) NHB2 <- NHB2 %>% split(NHB2$variable) foo <- NHB2 %>% map2(.,names(.),~select(.x,-variable)) %>% map2(.,names(.),~mutate(.x,lambda=as.numeric(gsub("lambda_","",.y)))) %>% map2(.,names(.),~mutate(.x,NHB="NHB")) %>% map2(.,names(.),~reshape2::dcast(.x,psa_id~NHB+strategy)) Optimal <- CEA <- list() for (i in names(foo)) { max.temp <- foo[[i]][,-1] %>% apply(.,1,max) Optimal[[i]] <- foo[[i]][,-1] %>% tbl_df() %>% mutate_all(funs(as.integer(.==max.temp))) CEA[[i]] <- colMeans(Optimal[[i]]) } CEA <- do.call("rbind",CEA) %>% tbl_df() %>% mutate(lambda=as.numeric(gsub("lambda_","",names(foo)))) colnames(CEA)<- gsub("NHB_","",colnames(CEA)) CEAC<-reshape2::melt(CEA, id.vars = "lambda") library(directlabels) txtsize<-12 CEAC <- CEAC %>% mutate(variable = paste0(" ",variable," ")) p <- ggplot(data = CEAC, aes(x = lambda, y = value, color = variable)) + geom_point() + geom_line() + #ggtitle("Cost-Effectiveness Acceptability Curves") + scale_colour_hue("Strategies: ",l=50) + #scale_x_continuous(breaks=number_ticks(6))+ xlab(expression("Policy Adoption Threshold "(lambda))) + ylab("Pr Cost-Effective") + theme_bw() + theme(legend.position="bottom",legend.title=element_text(size = txtsize), legend.key = element_rect(colour = "black"), legend.text = element_text(size = txtsize), title = element_text(face="bold", size=15), axis.title.x = element_text(face="bold", size=txtsize), axis.title.y = element_text(face="bold", size=txtsize), axis.text.y = element_text(size=txtsize), axis.text.x = element_text(size=txtsize))+scale_colour_grey(start = .5, end = 1)+ geom_dl(aes(label = variable), method = list(dl.combine( "last.points"), cex = 0.8)) return(p) } TornadoDiag <- function(indata,outcome,lambda) { # Get Outcome lhs <- indata %>% select(psa_id,contains("dQALY"),contains("dCOST")) %>% mutate(psa_id = row_number()) %>% reshape2::melt(id.vars='psa_id') %>% tidyr::separate(variable,c("outcome","strategy"),"_") %>% reshape2::dcast(psa_id+strategy~outcome) %>% mutate(NHB = dQALY-dCOST * lambda , NMB = dQALY*lambda - dCOST) # Get Parameters rhs <- indata %>% select(-contains("dQALY"),-contains("dCOST"), -contains("NMB"),-contains("NHB"),-psa_id) # Map to existing code inputs Strategies <- unique(lhs$strategy) Parms <- rhs %>% tbl_df() lhs$Y <- lhs[,outcome] Outcomes <- lhs %>% select(strategy,psa_id,Y) %>% reshape2::dcast(psa_id~strategy,value.var="Y") %>% select(-psa_id) # Find the Optimal opt<-which.max(colMeans(Outcomes)); opt # calculate min and max vectors of the parameters (e.g., lower 2.5% and 97.5%) X <- as.matrix(Parms) y <- as.matrix(Outcomes[,opt]) Y <- as.matrix(Outcomes) ymean <- mean(y) n <- nrow(Parms) nParams <- ncol(Parms) paramNames <- colnames(Parms) Parms.sorted <- apply(Parms,2,sort,decreasing=F) #Sort in increasing order each column of Parms lb <- 2.5 ub <- 97.5 Xmean <- rep(1,nParams) %*% t(colMeans(X)) XMin <- Xmean XMax <- Xmean paramMin <- as.vector(Parms.sorted[round(lb*n/100),]) paramMax <- as.vector(Parms.sorted[round(ub*n/100),]) diag(XMin) <- paramMin diag(XMax) <- paramMax XMin <- cbind(1, XMin) XMax <- cbind(1, XMax) X <- cbind(1,X) B <- solve(t(X) %*% X) %*% t(X) %*% y # Regression for optimal strategy library(matrixStats) bigBeta <- solve(t(X) %*% X) %*% t(X) %*% Y # Regression for all strategies yMin <- rowMaxs(XMin %*% bigBeta - ymean) yMax <- rowMaxs(XMax %*% bigBeta - ymean) ySize <- abs(yMax - yMin) rankY<- order(ySize) xmin <- min(c(yMin, yMax)) + ymean xmax <- max(c(yMin, yMax)) + ymean paramNames2 <- paste(paramNames, "[", round(paramMin,2), ",", round(paramMax,2), "]") strategyNames<-Strategies colfunc <- colorRampPalette(c("black", "white")) strategyColors <- colfunc(length(Strategies)) ## Polygon graphs: nRect <- 0 x1Rect <- NULL x2Rect <- NULL ylevel <- NULL colRect <- NULL for (p in 1:nParams){ xMean <- colMeans(X) xStart = paramMin[rankY[p]] xEnd = paramMax[rankY[p]] xStep = (xEnd-xStart)/1000 for (x in seq(xStart,xEnd, by = xStep)){ #for each point determine which one is the optimal strategy xMean[rankY[p] + 1] <- x # +1 moves beyond the constant yOutcomes <- xMean %*% bigBeta yOptOutcomes <- max(yOutcomes) yOpt <- strategyNames[which.max(yOutcomes)] if (x == xStart){ yOptOld <- strategyNames[which.max(yOutcomes)] y1 <- yOptOutcomes } #if yOpt changes, then plot a rectangle for that region if (yOpt != yOptOld | x == xEnd){ nRect <- nRect + 1 x1Rect[nRect] <- y1 x2Rect[nRect] <- yOptOutcomes ylevel[nRect] <- p colRect[nRect] <- yOptOld yOptOld <- yOpt y1 <- yOptOutcomes } } } txtsize <-8 d=data.frame(x1=x2Rect, x2=x1Rect, y1=ylevel-0.4, y2=ylevel+0.4, t=colRect, r = ylevel) p <- ggplot(d, aes(xmin = x1, xmax = x2, ymin = y1, ymax = y2, fill = t)) + xlab(paste0("Expected ",outcome)) + ylab("Parameters") + geom_rect()+ theme_bw() + scale_y_continuous(limits = c(0.5, nParams + 0.5),breaks=seq(1:ncol(Parms)), labels=paramNames2[rankY]) + scale_fill_grey(start = 0, end = .9)+ geom_vline(xintercept=ymean, linetype="dotted") + theme(legend.position="bottom",legend.title=element_text(size = txtsize), legend.key = element_rect(colour = "black"), legend.text = element_text(size = txtsize), title = element_text(face="bold", size=1), axis.title.x = element_text(face="bold", size=txtsize), axis.title.y = element_text(face="bold", size=txtsize), axis.text.y = element_text(size=txtsize), axis.text.x = element_text(size=txtsize))+ labs(fill="") return(p) } predict.ga <- function(object, n, n0, verbose = T){ #### Function to compute the preposterior for each of the #### basis functions of the GAM model. #### Inputs: #### - object: gam object #### - n: scalar or vector of new sample size to compute evsi on #### - n0: scalar or vector of effective prior sample size #### - verbose: Prints the variance reduction factor for each parameter ### Name of parameters names.data <- colnames(object$model) ### Create dataframe with parameter values data <- data.frame(object$model[,-1]) ## Name columns of dataframe colnames(data) <- names.data[-1] ### Number of parameters n.params <- ncol(data) ### Sanity checks if(!(length(n)==1 | length(n)==n.params)){ stop("Variable 'n' should be either a scalar or a vector the same size as the number of parameters") } if(!(length(n0)==1 | length(n0)==n.params)){ stop("Variable 'n0' should be either a scalar or a vector the same size as the number of parameters") } ### Make n & n0 consistent with the number of parameters if(length(n) == 1){ n <- rep(n, n.params) } if(length(n0) == 1){ n0 <- rep(n0, n.params) } ### Compute variance reduction factor v.ga <- sqrt(n/(n+n0)) if (verbose){ print(paste("Variance reduction factor =", round(v.ga, 3))) } ### Number of smoothers n.smooth <- length(object$smooth) ### Number of total basis functions n.colX <- length(object$coefficients) ### Number of observations n.rowX <- nrow(object$model) ### Initialize matrix for preposterior of total basis functions X <- matrix(NA, n.rowX, n.colX) X[, 1] <- 1 for (k in 1:n.smooth) { # k <- 1 klab <- substr(object$smooth[[k]]$label, 1, 1) if (klab == "s"){ Xfrag <- Predict.smooth.ga(object$smooth[[k]], data, v.ga[k]) } else { Xfrag <- Predict.matrix.tensor.smooth.ga(object$smooth[[k]], data, v.ga) } X[, object$smooth[[k]]$first.para:object$smooth[[k]]$last.para] <- Xfrag } ### Coefficients of GAM model Beta <- coef(object) ### Compute conditional Loss Ltilde <- X %*% Beta return(Ltilde) } Predict.smooth.ga <- function (object, data, v.ga = 1) { #### Function to compute the preposterior for each of the #### basis functions of a smooth for one parameter ### Produce basis functions for one parameter X <- PredictMat(object, data) # ‘mgcv’ version 1.8-17 ## Number of observations n.obs <- nrow(X) ### Apply variance reduction to compute the preposterior ### for each of the basis functions ## Vector of ones ones <- matrix(1, n.obs, 1) ## Compute phi on each of the basis function X.ga <- v.ga*X + (1-v.ga)*(ones %*% colMeans(X)) return(X.ga) } Predict.matrix.tensor.smooth.ga <- function (object, data, v.ga = rep(1, ncol(data))){ #### Function to compute the preposterior for each of the #### basis functions for one or more parameters and calculates #### the tensor product if more than one parameter is selected #### (Heavily based on function Predict.matrix.tensor.smooth from #### mgcv package) m <- length(object$margin) X <- list() for (i in 1:m) { # i <- 1 term <- object$margin[[i]]$term dat <- list() for (j in 1:length(term)) { # j <- 1 dat[[term[j]]] <- data[[term[j]]] } X[[i]] <- if (!is.null(object$mc[i])) # before: object$mc[i] PredictMat(object$margin[[i]], dat, n = length(dat[[1]])) # ‘mgcv’ version 1.8-17 else Predict.matrix(object$margin[[i]], dat) n.obs <- nrow(X[[i]]) } # end for 'i' mxp <- length(object$XP) if (mxp > 0) for (i in 1:mxp) if (!is.null(object$XP[[i]])) X[[i]] <- X[[i]] %*% object$XP[[i]] ### Apply variance reduction to compute the preposterior ### for each of the basis functions ## Vector of ones ones <- matrix(1, n.obs, 1) ## Initialize and fill list with preposterior of basis functions ## for each parameter X.ga <- list() for (i in 1:m) { # i <- 1 X.ga[[i]] <- v.ga[i]*X[[i]] + (1-v.ga[i])*(ones %*% colMeans(X[[i]])) } ### Compute tensor product T.ga <- tensor.prod.model.matrix(X.ga) # ‘mgcv’ version 1.8-17 return(T.ga) } ## For Simulating Medicaid cov_sim <- function(params) { p <- params[grep("^p_",names(params))] %>% unlist() R <- params[grep("^R_",names(params))] %>% unlist() %>% data.frame() %>% rownames_to_column(var = "type") %>% separate(type,into= c("exa","exp"), sep ="_TO_") %>% set_names(c("exa","exp","value")) %>% spread(exp,value) %>% select(-exa) %>% as.matrix() DD <- params[grep("^DD_",names(params))] %>% unlist() %>% data.frame() %>% rownames_to_column(var = "type") %>% separate(type,into= c("exa","exp"), sep ="_TO_") %>% set_names(c("exa","exp","value")) %>% spread(exp,value) %>% select(-exa) %>% as.matrix() baseline <- t(p) %*% R expmedicaid <- t(p) %*% (R + DD) mvpf_med <- calculate_wtp_public(params) mvpf_subsidy <- simulate_subsidy(params) R_subsidy <- R R_subsidy[4,2] <- R_subsidy[4,4] * params$frac_uninsured_elig * mvpf_subsidy$takeup R_subsidy[4,4] <- 1 - sum(R_subsidy[4,1:3]) subsidy <- t(p) %*% R_subsidy # Now need to include estimation of MVPF cost and benefits. out <- list(baseline = baseline, med = expmedicaid, subsidy = subsidy) %>% bind_rows() %>% tbl_df() %>% mutate(type = insurance_sipp_lut) %>% select(type,baseline,med, subsidy) %>% gather(key,value,-type) %>% mutate(iteration = 1) %>% unite("tmp",key,type) %>% spread(tmp,value) %>% bind_cols(mvpf_med %>% data.frame()) %>% rename(med_mvpf = mvpf, med_mvpf_num = mvpf_num, med_mvpf_denom = mvpf_denom, med_wtp = wtp, med_cost = cost, med_N = N) %>% bind_cols(mvpf_subsidy %>% data.frame()) %>% rename(subsidy_takeup = takeup, subsidy_C_H = C_H, subsidy_uncomp = uncomp, subsidy_mvpf_num = mvpf_num, subsidy_mvpf_denom = mvpf_denom, subsidy_mvpf = mvpf) return(out) } calculate_wtp_public <- function(params, scaling_factor = 1) { #The net cost of Medicaid equals the average increase in medical spending due to Medicaid # plus the average decrease in out-of-pocket spenign due to Medicaid (see equation 22). p_1 <- params$OOP_Tx / params$G p_0 <- params$OOP_Cx / params$G_Cx MCD_SPEND = params$G - params$G_Cx C = MCD_SPEND + params$OOP_Cx # The monetary transfer from Medicaid to external parties, N, is the difference between G and C. N <- params$G - C welfare_weight <- params$v_i / params$v_j # We estimate the transfer component and pure-insurance component separately, and combine them # for our estimate of \gamma(1). # Transfer component (p.29) # using a linear approximation and the estimates of E[m(0,\theta)] and E[m(1,\theta)] Tr <- (p_0-p_1)*(0.5 * (params$G_Cx + params$G)) net_cost_as_frac_gross <- C / params$G moral_hazard_cost <- params$G - Tr - N # wtp <- Tr + params$I # # mvpf_gov <- wtp / C # mvpf_indiv <- (wtp + params$G * welfare_weight * (N / params$G)) / params$G # # mvpf_num_gov <- wtp # mvpf_denom_gov <- C # mvpf_num_indiv <- (wtp + params$G * welfare_weight * (N / params$G)) # mvpf_denom_indiv <- params$G # # mvpf_num <- params$gov_incidence * mvpf_num_gov + (1 - params$gov_incidence) * mvpf_num_indiv # mvpf_denom <- params$gov_incidence * mvpf_denom_gov + (1 - params$gov_incidence) * mvpf_denom_indiv # # mvpf <- mvpf_num / mvpf_denom #out <- list(mvpf = mvpf , mvpf_num = (mvpf_num / 12)/100 , mvpf_denom = (mvpf_denom / 12)/100, wtp = wtp , cost = C , N = N) ###################################################### # Scale all relevant values by government cost of # Medicaid so it can be measured in terms of a single # dollar spent on Medicaid. ###################################################### scaling_factor = C wtp <- Tr + params$I + params$fudge mvpf_gov <- (wtp/scaling_factor ) / (C/scaling_factor ) mvpf_indiv <- (wtp/scaling_factor + params$G/scaling_factor * welfare_weight * (N / params$G)) / (params$G/scaling_factor ) mvpf_num_gov <- wtp / scaling_factor mvpf_denom_gov <- C /scaling_factor mvpf_num_indiv <- (wtp/scaling_factor + params$G/scaling_factor * welfare_weight * (N / params$G)) mvpf_denom_indiv <- params$G/scaling_factor mvpf_num <- params$gov_incidence * mvpf_num_gov + (1 - params$gov_incidence) * mvpf_num_indiv mvpf_denom <- params$gov_incidence * mvpf_denom_gov + (1 - params$gov_incidence) * mvpf_denom_indiv mvpf <- mvpf_num / mvpf_denom out <- list(mvpf = mvpf , mvpf_num = mvpf_num , mvpf_denom = mvpf_denom, wtp = wtp , cost = C , N = N) return(out) } simulate_subsidy <- function(params) { takeup <- get_takeup(params, premium = params$plan_premium) # takeup_deriv = 1/(get_takeup(params, premium = params$plan_premium+1) - get_takeup(params, premium = params$plan_premium)) # cost_reformed <- get_cost(params, premium = params$plan_premium) # uncomp = fn_uncomp(cost = cost_reformed, uninsured_oop_share = params$uninsured_oop_share, phi = params$phi) # # mvpf_num = takeup + # params$eta * # ( # pmax(0,uncomp) / # (-1 * takeup_deriv) # ) # mvpf_denom = takeup + # ((pmax(0,cost_reformed - params$gov_incidence * uncomp - params$plan_premium)) / # (-1 * takeup_deriv)) # mvpf = mvpf_num / mvpf_denom # # Alternative Version s_star <- get_takeup(params, premium = params$plan_premium) ds_dpH <- get_takeup(params, premium = params$plan_premium)-get_takeup(params, premium = params$plan_premium+1) C_H <- get_cost(params, premium = params$plan_premium) p_H <- params$plan_premium uncomp = fn_uncomp(cost = C_H, uninsured_oop_share = params$uninsured_oop_share, phi = params$phi) welfare_weight <- params$v_i / params$v_j mvpf_num <- s_star + welfare_weight * uncomp * ds_dpH cost_of_new_enrollees <- ds_dpH * (C_H - params$gov_incidence * uncomp - p_H) mvpf_denom <- s_star + cost_of_new_enrollees mvpf<- mvpf_num / mvpf_denom output <- list( takeup = takeup, # takeup_deriv = takeup_deriv, C_H = C_H, uncomp = uncomp, mvpf_num = mvpf_num, mvpf_denom = mvpf_denom, mvpf = mvpf ) return(output) }
0853dd67327248b42904eac0f4b478575d4818be
ce9f7ef440ba0a08d07e7a181fead89fd5874615
/featuresToBed.R
4088f58dd21559904d764ce40e253f4965d9bb1b
[ "CC0-1.0" ]
permissive
karakulahg/DTExpress
706c7ce768a616fa92ad34cfe4499f860c4d907f
b344f752dd71ec3cf7df29644bfc05a4a67eea48
refs/heads/main
2023-06-14T04:16:36.577983
2023-04-25T09:41:30
2023-04-25T09:41:30
200,231,047
0
0
null
null
null
null
UTF-8
R
false
false
732
r
featuresToBed.R
args = commandArgs(trailingOnly = TRUE) if (length(args) < 1) { stop("At least one argument must be supplied (input file).n", call. = FALSE) } else if (length(args) == 1) { # default output file args[2] = "output.bed" } options(stringsAsFactors = F) #read.csv("GCF_000001735.4_TAIR10.1_feature_table.txt", sep = "\t") -> features read.csv(args[1], sep = "\t") -> features features.subset <- subset(features, X..feature == "gene") which(features.subset$symbol == "") -> symbolMissing features.subset$symbol[symbolMissing] <- features.subset$locus_tag[symbolMissing] features.bed <- features.subset[, c(7, 8, 9, 15, 16, 10)] write.table( features.bed, args[2], quote = F, col.names = F, row.names = F, sep = "\t" )
9e795760b7b95d970368689182bb6a68b82027c2
6840c1a88a8aa9bbe2723c9d424c3126751bca7f
/man/check_prior.Rd
13599f097465632554f32e3965b2ba346af848ef
[]
no_license
brieuclehmann/multibergm
edd8170bc5dc7892254974d4aeb979f4c0fe5890
a5f57abc8da185534f38eae9b84a0c58573931d1
refs/heads/master
2022-06-11T14:38:53.370619
2022-06-01T08:18:40
2022-06-01T08:18:40
206,797,419
3
0
null
2021-05-12T12:59:16
2019-09-06T13:15:28
R
UTF-8
R
false
true
811
rd
check_prior.Rd
% Generated by roxygen2: do not edit by hand % Please edit documentation in R/set_priors.R \name{check_prior} \alias{check_prior} \alias{check_prior_mean} \alias{is_covmat} \alias{check_prior_cov} \alias{check_prior_scale} \alias{check_prior_df} \title{Check validity of multibergm prior} \usage{ check_prior(prior, n_terms, n_groups) check_prior_mean(x, n_terms) is_covmat(x) check_prior_cov(x, n_terms) check_prior_scale(x, n_terms) check_prior_df(x, n_terms) } \arguments{ \item{prior}{A list of explicit prior specifications.} \item{n_terms}{Number of terms (summary statistics) in the exponential random graph model} \item{n_groups}{Number of distinct groups} \item{x}{Prior mean or covariance to be checked} } \description{ Internal functions to check compatibility of the prior with the model. }
1729bd0497be9add6ab34a0f7e9107414c6e47d0
7f72ac13d08fa64bfd8ac00f44784fef6060fec3
/RGtk2/man/gtkFontButtonGetShowStyle.Rd
112e19762307e992b097cffa28ee2ba0d06658e5
[]
no_license
lawremi/RGtk2
d2412ccedf2d2bc12888618b42486f7e9cceee43
eb315232f75c3bed73bae9584510018293ba6b83
refs/heads/master
2023-03-05T01:13:14.484107
2023-02-25T15:19:06
2023-02-25T15:20:41
2,554,865
14
9
null
2023-02-06T21:28:56
2011-10-11T11:50:22
R
UTF-8
R
false
false
452
rd
gtkFontButtonGetShowStyle.Rd
\alias{gtkFontButtonGetShowStyle} \name{gtkFontButtonGetShowStyle} \title{gtkFontButtonGetShowStyle} \description{Returns whether the name of the font style will be shown in the label.} \usage{gtkFontButtonGetShowStyle(object)} \arguments{\item{\verb{object}}{a \code{\link{GtkFontButton}}}} \details{Since 2.4} \value{[logical] whether the font style will be shown in the label.} \author{Derived by RGtkGen from GTK+ documentation} \keyword{internal}
086e77f7b14e6f82f64c3c0d2739e67dfba9e3c6
02f3c8b878eb6e8d2a99c547f251b3a4dd2805da
/man/grib_cube.Rd
06a963303d59c88e146bea04350bf275ea69b9cc
[ "BSD-3-Clause" ]
permissive
nawendt/gribr
86d4d7f228b12f4dd44653e8e06cdb1d42446d73
3d2a6f438a1b88c36cb5071906ab2a5ddbff1041
refs/heads/main
2023-07-19T15:06:39.424009
2023-07-10T14:13:44
2023-07-10T14:13:44
51,981,861
20
1
BSD-3-Clause
2023-04-03T14:22:37
2016-02-18T05:32:30
R
UTF-8
R
false
true
1,621
rd
grib_cube.Rd
% Generated by roxygen2: do not edit by hand % Please edit documentation in R/grib_cube.R \name{grib_cube} \alias{grib_cube} \title{Create 3D volume of a GRIB variable} \usage{ grib_cube(gribObj, shortName, typeOfLevel, decreasing = FALSE) } \arguments{ \item{gribObj}{\code{GRIB} class object.} \item{shortName}{The short name given in the GRIB file of the variable to select.} \item{typeOfLevel}{The vertical coordinate to use as given by the typeOfLevel key in the GRIB file.} \item{decreasing}{Parameter to tell the array's vertical coordinate to be increasing or decreasing.} } \value{ Returns a three-dimenional array. } \description{ \code{grib_cube} creates a three-dimensional array from one variable along a chosen vertical coordinate. } \details{ \code{grib_cube} is a wrapper function for \code{grib_select} to conveniently create a three-dimensional cube. The user inputs a variable to search for and the vertical coordinate to use when finding each level. Because \code{grib_cube} uses \code{grib_select}, speed can become an issue. This is meant as a convenience to "get the job done". If you want more speed, it will always be better to know which message number you want, set up your own loop, and use \code{grib_get_message} as that will avoid the overhead of searching through the GRIB file. } \examples{ g <- grib_open(system.file("extdata", "lfpw.grib1", package = "gribr")) cube <- grib_cube(g, 'u', 'isobaricInhPa', TRUE) grib_close(g) } \seealso{ \code{\link{grib_get_message}} \code{\link{grib_list}} \code{\link{grib_expand_grids}} \code{\link{grib_latlons}} \code{\link{grib_select}} }
94fec547bb361970a5b5c2177f32bf2eb9e725c8
ae3173d4ffc3e1e9f45acd00772a30760ec2d187
/mtDNAheatmap.R
9b96ca648de0f19d2916c21d4c0e70fa3baec358
[]
no_license
cmiciano/RNA-seq-FA18
41d854ae9fb8cf10cbaf40fbaca1521745088b81
2983653f50a60b82ef05e267e50c9fd2ff9eaa18
refs/heads/master
2022-02-21T19:57:32.081617
2019-10-04T18:31:58
2019-10-04T18:31:58
212,876,971
0
0
null
null
null
null
UTF-8
R
false
false
3,174
r
mtDNAheatmap.R
library(gplots) yb<-colorRampPalette(c("blue","white","red"))(100) fpkmyeast <- read.delim("~/Documents/salk/fpkmyeast.txt") #Change rownames of fpkm to standard gene symbol names(fpkmyeast)[1]<-"Transcript" fpkmname <- as.character(fpkmyeast$Annotation.Divergence) fpkmsym<- sapply(fpkmname, function(x) strsplit(x,"|", fixed = T)[[1]][1]) rownames(fpkmyeast) <- fpkmsym #Read in curated and high throughput curatedgenes <- read.table("~/Downloads/mitochondrion_annotations_Manually Curated.txt", sep = "\t", skip = 1, header = T, comment.char = "!") highthrugenes <- read.table("~/Downloads/mitochondrion_annotations_High-throughput.txt", sep = "\t", skip = 1, header = T, comment.char = "!") #Write curated genes for IPA pathway analysis transcur <- as.character(curatedgenes$Gene.Systematic.Name) curinfpkm <- subset(fpkmyeast, fpkmyeast$Transcript %in% transcur) curtransf <- log(curinfpkm[,9:14]+5) #genes by sample #Remove genes that haven't changed between WT and KO curvar <- curtransf[ apply(curtransf, 1, var, na.rm = TRUE) != 0 , ] #Renamed curated col names for heatmap comb2cur <- sapply(colnames(curvar), function(x) paste(strsplit(x,"_", fixed = T)[[1]][2], strsplit(x,"_",fixed = T)[[1]][3], sep = "_" )) colnames(curvar) <- comb2cur #############Curated heatmap ############## pdf("Curated.pdf", width=7, height=7) par(mar=c(2,2,2,2), cex=1.0) heatmap.2(as.matrix(curvar), col=yb, scale="row", dendrogram = "row", labRow = "", key=TRUE, symkey=FALSE, density.info="none", trace="none", cexRow=0.3,lhei=c(0.20,0.70), lwid = c(0.25,0.5), reorderfun=function(d,w) reorder(d, w, agglo.FUN=mean), distfun=function(x) as.dist(1-cor(t(x))), hclustfun=function(x) hclust(x, method="ward.D2")) dev.off() #####High throughput genes transhighthru <- as.character(highthrugenes$Gene.Systematic.Name) highinfpkm <- subset(fpkmyeast, fpkmyeast$Transcript %in% transhighthru) hightransf <- log(highinfpkm[,9:14]+5) #genes by sample #Remove genes that haven't changed between WT and KO highvar <- hightransf[ apply(hightransf, 1, var, na.rm = TRUE) != 0 , ] #Renamed curated col names for heatmap comb2high <- sapply(colnames(highvar), function(x) paste(strsplit(x,"_", fixed = T)[[1]][2], strsplit(x,"_",fixed = T)[[1]][3], sep = "_" )) colnames(highvar) <- comb2high #############High throughput heatmap ############## pdf("Highthroughput.pdf", width=7, height=7) par(mar=c(2,2,2,2), cex=1.0) heatmap.2(as.matrix(highvar), col=yb, scale="row", dendrogram = "row", labRow = "", key=TRUE, symkey=FALSE, density.info="none", trace="none", cexRow=0.3,lhei=c(0.20,0.70), lwid = c(0.25,0.5), reorderfun=function(d,w) reorder(d, w, agglo.FUN=mean), distfun=function(x) as.dist(1-cor(t(x))), hclustfun=function(x) hclust(x, method="ward.D2")) dev.off()
792936326bd5a3c84fb035aaa9ce78d9b4a591c7
091518481f7e2d022c98ea16a45eb3491953671e
/R/assignECMWFTDataToMonitor.R
e0d7f091ebeae2771ca94f765d6dd2319ae79af4
[ "MIT" ]
permissive
stevenjoelbrey/SmokeInTheCity
87625532edf1721a7deb5ab183d159b44d964e36
481a4d2b70fe380e14fe75e8e4fdecbf6b41001f
refs/heads/master
2021-01-10T03:28:58.049008
2017-01-24T18:48:40
2017-01-24T18:48:40
50,605,479
0
2
null
null
null
null
UTF-8
R
false
false
10,194
r
assignECMWFTDataToMonitor.R
# assignReanalysisTDataToMonitor.R # This script assigns 00Z ecmwf reanalysis data to monitors # 2m - ecmwf data was downloaded # http://apps.ecmwf.int/datasets/data/interim-full-daily/levtype=sfc/ library(fields) library(maps) library(geosphere) ################################################################################ # Assigns gridded temperature data to PM monitors passed here ################################################################################ getStationWeatherData <- function(workSpaceData=workSpaceData, sanityCheck=FALSE){ # Load the ecmwf Temperature data load("ecmwfData/t2m.RData") gridLon <- t2m[["gridLon"]] gridLat <- t2m[["gridLat"]] ecmwfTime <- t2m[["ecmwfTime"]] t2m <- t2m[["t2m"]] # this will erase the list # Plot the data on a map, with monitor locations, to make sure everything # lines up and makes sense map('world') image.plot(gridLon, gridLat,t2m[,,1], add=TRUE) map('world', add=TRUE) # Load the monitor locations to plot with this data, do they overlap? lon <- workSpaceData[['lon']] lat <- workSpaceData[['lat']] points(lon,lat, pch=19, cex=0.5) image.plot(gridLon, gridLat,t2m[,,1]) map("world", xlim=c(min(gridLon), max(gridLon)), ylim=c(min(gridLat), max(gridLat)), add=TRUE) map('state', add=TRUE) points(lon,lat, pch=19) title(paste("showing the ecmwf data grid for ", ecmwfTime[1])) # Load the cloud cover data (exact same grid, same lon lat and time) load("ecmwfData/tcc.RData") tcc <- tcc[["tcc"]] # this will erase the list ############################################################################## # NOTE ON ECMWFTime: ############################################################################## # This 00Z time is 6:00 PM MDT 8PM EDT and 5 PM PDT of the PREVIOUS day. # 00Z is the date of the day just starting, still previous date in US at this # hour. # AKA: 00Z is 7 hours ahead of time here in Fort Collins MST. # http://scc-ares-races.org/generalinfo/utcchart.html # http://www.timeanddate.com/worldclock/timezone/zulu # So the date in North America is not the same as that of 00Z because they are # 7 hours ahead and 00Z has the date of the day just starting. # Take a day away to make this time more useful to the PM and ozone data # which is in US time and date. That daily data will most meaningfully be # matched with evenning temperature data of the 00Z data. # NOTE: Evening temperature snapshot is not a perfect way to determine # NOTE: ozone relevant ozone production but should be decent at determining # NOTE: what days are generally warmer and cooler than others. secondsInDay <- 24 * 60^2 ecmwfTimeModified <- ecmwfTime - secondsInDay # Load requested data packet Hybrid_mdf <- workSpaceData[["Hybrid_mdf"]] PM_df <- workSpaceData[["PM_df"]] lon <- workSpaceData[["lon"]] lat <- workSpaceData[["lat"]] # copycat will be replaced with T and cloud fraction values from ecmwf T_df <- PM_df # temperature dataframe CC_df <- PM_df # cloud cover dataframe nMonitors <- dim(PM_df)[2] # columns of PM_df are PM Monitors at stations # These dates are made from local times in the US. which are a day behind 00Z # I am placing them in UTC time zone so that it will be possible to match # the modified ecmwf time series. In reality there is no actual time of day # associated with daily PM and ozone data. measuredDataDate <- rownames(T_df) PMTime <- as.POSIXct(measuredDataDate, tz="UTC") # SO THAT WE CAN MATCH! # NOTE: The time zone of the loaded ecmwf time is UTC. # Loop through monitors and assign the appropriate grid of ecmwf time series # data for (i in 1:nMonitors){ # Clear out the dataframe of PM values resulting from copying T_df[,i] <- NA CC_df[,i] <- NA # Find the grid box the monitor falls inside of, create a mask lonIndex <- which.min(abs(gridLon - lon[i])) latIndex <- which.min(abs(gridLat - lat[i])) # Compute the haversine distance of monitor to gridpoint center chosen meterPerkm <- 1000 glon <- gridLon[lonIndex] glat <- gridLat[latIndex] greatCircleDistance <- distHaversine(c(glon,glat), c(lon[i], lat[i]))/meterPerkm #print(greatCircleDistance) # The maximum dy of this ecmwf data should be the hypotenouse of the # A=dy/2, B=dy/2 of ecmwf data (dy=0.75 degrees). C = (111.3195/2km^2 + 111.3195/2^2)^.5 # so C=78.71477km maxDistanceAccepted <- 64 #km if(greatCircleDistance > maxDistanceAccepted){ stop("You are not pulling the correct temperature data for this monitor.") } # Plot the grid center and monitor location, contributes to general sanity points(lon[i],lat[i], col="white", pch=19) points(gridLon[lonIndex], gridLat[latIndex], col="pink", pch=19) points(gridLon[lonIndex], gridLat[latIndex], col="pink", pch=3) # Exstract monitors temperature array temp <- t2m[lonIndex, latIndex,] # Convert to degrees C temp <- temp - 273.15 # 0C == 273.15 K # Now convert the temperature from C to f temp <- temp * 9/5 + 32.0 # Exstract monitors cloud cover fraction array CC <- tcc[lonIndex, latIndex,] # Where in ecmwfTimeModified do PMTimes land? matchecmwfTimeToPM <- match(PMTime, ecmwfTimeModified) T_df[,i] <- temp[matchecmwfTimeToPM] CC_df[,i] <- CC[matchecmwfTimeToPM] # Commented date assigment in lines below is the most effective way to make # sure that you are assigning the correct dates temperature data using # the match() function. This has caused major headaches before... ##T_df[,i] <- as.character(ecmwfTimeModified[matchecmwfTimeToPM]) ##CC_df[,i] <- as.character(ecmwfTimeModified[matchecmwfTimeToPM]) } # End of for loop looping through PM/ozone monitors # Place new variables into the workspace workSpaceData[["T_df"]] <- T_df workSpaceData[["CC_df"]] <- CC_df print("assigned temperature and cloud data based on ecmwf grid") return(workSpaceData) } # ################################################################################ # # Now use this ecmwf temperature data to create a clear T mask that ensures # # smoke-free days are warmer than smoke-impacted days. # ################################################################################ # createSmokeFreeTMask <- function(workSpaceData, # TSdFactor = 1, # applySkyMask=FALSE, # maxCloudCoverPercent=10){ # # # Get the required data # T_df <- workSpaceData[["T_df"]] # smokeImpactMask <- workSpaceData[["smokeImpactMask"]] # # # set dimensions for new temperature mask # nMonitor <- dim(T_df)[2] # meanSmokedT <- rep(NA, nMonitor) # sdSmokedT <- rep(NA, nMonitor) # smokeFreeMask <- smokeImpactMask # Copying for proper dimensions and labels # # # Loop through each monitor, figuring out the temperature threshold based # # on arguments given to this function # for (i in 1:nMonitor){ # # # clear out smokeImpactMask data # smokeFreeMask[,i] <- FALSE # Assume FALSE until proven otherwise # # # Which rows are smoke-impacted based on work so far? # smokedRows <- smokeImpactMask[,i] # smokedDaysT <- as.numeric(T_df[smokedRows,i]) # # # Get the statistics on the smoke-impacted temperatures # meanSmokedT[i] <- mean(smokedDaysT, na.rm=TRUE) # sdSmokedT[i] <- sd(smokedDaysT, na.rm=TRUE) # # # Figure out where the temperature is greater than smoky day average # TThresh <- meanSmokedT[i] + sdSmokedT[i] * TSdFactor # TCuttoffMask <- T_df[,i] >= TThresh # # # if(is.nan(TThresh) & sum(smokedRows)==0){ # # There are no smoke impacted days, so all rows are smoke free. # # We know this because TThresh is not a number and there are zero smoked # # rows. # smokeFreeRows <- rep(TRUE, length(smokedRows)) # } else { # # There are smoke impacted days so we need to choose smoke-free carefully # # Also, we want to be sure that these warm days are not also smoke-impacted! # # NOTE: smokedRows == TRUE where smoke-impacted. Use ! to change those to # # NOTE: FALSE and smoke-free days to TRUE # smokeFreeRows <- !smokedRows & TCuttoffMask # # # TODO: Could add PM mask as well to ensure that clear days are not high # # TODO: PM days. Require PM measurement for smokeFreeDays # # } # # Store the smokeFreeRows (days) information in TMask # smokeFreeMask[,i] <- smokeFreeRows # # } # # # Include the smokeFreeMask in the workspace data # workSpaceData[["smokeFreeMask"]] <- smokeFreeMask # workSpaceData[["meanSmokedT"]] <- meanSmokedT # workSpaceData[["sdSmokedT"]] <- sdSmokedT # # # For testing purposes give this information back to console # print(paste("The sum of smokeFreeMask after T-Control is:", # sum(smokeFreeMask,na.rm=TRUE))) # # # # Now apply the skycondition mask if desired # if(applySkyMask){ # # # Get the Cloud Cover Dataframe # CC_df <- workSpaceData[["CC_df"]] * 100 # to make % # # # Where are the skies more clear than specified %? # cloudFreeMask <- CC_df <= maxCloudCoverPercent # # # Add this cloudMask to the workspace # workSpaceData[["cloudFreeMask"]] <- cloudFreeMask # # # Modify smokeFreeMask based on this new cloud condition # smokeFreeMaskNew <- smokeFreeMask & cloudFreeMask # # # Overwrite the original smokeFreeMask # workSpaceData[["smokeFreeMask"]] <- smokeFreeMaskNew # # # print(paste("The sum of smokeFreeMask after sky-control is:", # sum(smokeFreeMaskNew,na.rm=TRUE))) # print("If the later is not small than the former, you have a problem.") # } # # # Return the appended workSpaceData # return(workSpaceData) # # # }
7b6c98a65a6f7db03608d035b4b91063e086ab54
109734b597c2d760725a1a050174a5d11b3c1a9b
/man/Extract.anylist.Rd
ebeb85fb7fc71c3890426e92727c8354d4e15828
[]
no_license
rubak/spatstat
c293e16b17cfeba3e1a24cd971b313c47ad89906
93e54a8fd8276c9a17123466638c271a8690d12c
refs/heads/master
2020-12-07T00:54:32.178710
2020-11-06T22:51:20
2020-11-06T22:51:20
44,497,738
2
0
null
2020-11-06T22:51:21
2015-10-18T21:40:26
R
UTF-8
R
false
false
1,206
rd
Extract.anylist.Rd
\name{Extract.anylist} \alias{[.anylist} \alias{[<-.anylist} \title{Extract or Replace Subset of a List of Things} \description{ Extract or replace a subset of a list of things. } \usage{ \method{[}{anylist}(x, i, \dots) \method{[}{anylist}(x, i) <- value } \arguments{ \item{x}{ An object of class \code{"anylist"} representing a list of things. } \item{i}{ Subset index. Any valid subset index in the usual \R sense. } \item{value}{ Replacement value for the subset. } \item{\dots}{Ignored.} } \value{ Another object of class \code{"anylist"}. } \details{ These are the methods for extracting and replacing subsets for the class \code{"anylist"}. The argument \code{x} should be an object of class \code{"anylist"} representing a list of things. See \code{\link{anylist}}. The method replaces a designated subset of \code{x}, and returns an object of class \code{"anylist"}. } \seealso{ \code{\link{anylist}}, \code{\link{plot.anylist}}, \code{\link{summary.anylist}} } \examples{ x <- anylist(A=runif(10), B=runif(10), C=runif(10)) x[1] <- list(A=rnorm(10)) } \author{ \spatstatAuthors } \keyword{spatial} \keyword{list} \keyword{manip}
052ad634c60461990965f8ea791d9f2cc29572a3
ae38e72efce94388334551b1864c297a3824ec3a
/man/create_airly_api_response.Rd
687d6da73408e67cad860cf65460864fcda16a47
[]
no_license
cran/aiRly
dde7a8a722247b69fd3ad1b5b4d08e0d45614692
a950aff2eab1389e7424dacee07e6a499746e77c
refs/heads/master
2021-04-05T11:17:24.630452
2020-03-19T13:00:02
2020-03-19T13:00:02
248,550,766
0
0
null
null
null
null
UTF-8
R
false
true
697
rd
create_airly_api_response.Rd
% Generated by roxygen2: do not edit by hand % Please edit documentation in R/airly_api_response.R \name{create_airly_api_response} \alias{create_airly_api_response} \title{Creates an object representing a response from the Airly API. Also every API call return information about current limits What is used to assign variables in pkg.env} \usage{ create_airly_api_response(response) } \arguments{ \item{response}{response object} } \value{ object representing a response from the Airly API } \description{ Creates an object representing a response from the Airly API. Also every API call return information about current limits What is used to assign variables in pkg.env }
b8940ad2b088b249bce9bce1105ea1cc936b35fa
f7114c3a1ee36fee6fefebeb468c7ad0131cac46
/fire.pcvalues.R
2a9eff0dab1e5a2b0844a09d8e30462b9eaac203
[]
no_license
HaireLab/Fuego-en-la-Frontera
6058639da782349536a98723b41e3d6a8b2a37f4
15b34fef4779c549b9ca86475e1a59c349670456
refs/heads/master
2023-09-03T12:39:57.239902
2021-10-27T13:56:09
2021-10-27T13:56:09
279,133,955
1
0
null
null
null
null
UTF-8
R
false
false
2,293
r
fire.pcvalues.R
##---------------------------------------------------------------- ## ## fire.polygon.pcvalues.R ## ## Purpose: append princiipal component values (mean within fire perimeter) ## to fire polygon shp to use in analysis of high severity metrics ## ## Author: S. Haire, @HaireLab ## ## ## Date: 3 july 2020 ## ##-------------------------------------------------------------- ## ## Notes: ## Before running this script, download the fire perimeters ## < https://doi.org/10.5066/P9BB5TIO > ## ## PC1 and PC2 rasters are available in the respository data folder ## ##--------------------------------------------------------------- library(raster) library(rgdal) library(landscapemetrics) library(plyr) library(dplyr) library(geosphere) ## Read in the data and project the fire perimeter polygons to match the principal component layers ## paths to input data perimpath<-'./data/sp'# fire perimeters...put the shp with new attributes here too pcpath<-'./data/PCA/' ## PCA rasters ## data ## pc's have bioclim data projection bioclim.prj<-"+proj=lcc +lat_1=49 +lat_2=77 +lat_0=0 +lon_0=-95 +x_0=0 +y_0=0 +datum=WGS84 +units=m +no_defs" # from metadata lambert conformal conic ## read in pc's and assign prj pc1<-raster(paste(pcpath, "PC1b.rot.tif", sep="")); crs(pc1)<-bioclim.prj pc2<-raster(paste(pcpath, "PC2b.rot.tif", sep="")); crs(pc2)<-bioclim.prj ## read in perimeters/sp polys and project the perimeters to match the pc's perims<-readOGR(perimpath, "Sky_Island_Fire_Polys_1985_2017") perims.lcc<-spTransform(perims, bioclim.prj) ## Extract pc 1 & pc2 values and output the mean w/in polygon (fire perimeter). Save appended shp. ## stack the pc's and extract values within the polygons s<-stack(pc1,pc2) pc.ex<-extract(s, perims.lcc, method="bilinear",df=TRUE) ## calulate the mean values within the fire perimeters mean.pc<-ddply(pc.ex,~ID, summarise, mean.pc1 = mean(PC1b.rot), mean.pc2=mean(PC2b.rot)) ## add pc mean values to the spatial polygons perims$pc1<-mean.pc[,2]; perims$pc2<-mean.pc[,3] perims<-perims[,c(1:5,15,16)] ## just save the year, name & id, country, sky island and pc values ## save the perim shp writeOGR(perims, "./data/sp", "fire.polys.pcvalues", driver="ESRI Shapefile", overwrite=TRUE)
36a3356a3fb5056285f0c46868f1d82e88c68f2b
7a031aad2eba871dd062de1b5a1b790c5cb64ea5
/R/fwiGrid.R
6de984e871556f4de611e2061fcdf2aa2c80102a
[]
no_license
SantanderMetGroup/fireDanger
7554a254dbbbe51f813f012beb184b95ef4815ee
a2cbeb780184b695d58f37565258c423f2249931
refs/heads/master
2020-12-25T03:00:52.506116
2018-10-04T15:04:21
2018-10-04T15:04:21
61,045,602
1
1
null
2016-06-13T15:02:27
2016-06-13T15:02:27
null
UTF-8
R
false
false
14,069
r
fwiGrid.R
#' @title Fire Weather Index applied to multigrids #' #' @description Implementation of the Canadian Fire Weather Index System for multigrids #' #' @param multigrid containing Tm (temperature records in deg. Celsius); H (relative humidity records in \%); #' r (last 24-h accumulated precipitation in mm); W (wind velocity records in Km/h). See details. #' @param mask Optional. Binary grid (0 and 1, 0 for sea areas) with \code{dimensions} attribute \code{c("lat", "lon")}. #' @param what Character string. What component of the FWI system is computed?. Default to \code{"FWI"}. #' Note that unlike \code{\link{fwi1D}}, only one single component choice is possible in \code{fwiGrid}. #' See \code{\link{fwi1D}} for details and possible values. #' @param nlat.chunks For an efficient memory usage, the computation of FWI can be split into #' latitudinal bands (chunks) sepparately. The number of chunks is controlled here. #' Default to \code{NULL} (i.e., no chunking applied). #' @param restart.annual Logical. Should the calculation be restarted at the beginning of every year? #' If the grid encompasses just one season (e.g. JJA...), this is the recommended option. Default to \code{TRUE}. #' @param ... Further arguments passed to \code{\link{fwi1D}}. #' @template templateParallelParams #' #' @return A grid corresponding to the variable defined in argument \code{what}. #' #' @details #' #' \strong{Variable names} #' #' The variables composing the input multigrid are expected to have standard names, as defined by the dictionary #' (their names are stored in the \code{multigrid$Variable$varName} component). #' These are: \code{"tas"} for temperature, \code{"tp"} for precipitation, \code{"wss"} for windspeed. In the case of relative humidity, #' either \code{"hurs"} or \code{"hursmin"} are accepted, the latter in case of FWI calculations according to the \dQuote{proxy} version #' described in Bedia \emph{et al} 2014. #' #' Note that the order of the variables within the multigrid is not relevant. These are indexed by variable names. #' #' \strong{Landmask definition} #' #' The use of a landsmask is highly recommended when using RCM/GCM data becasue (i) there is no point in calculating #' FWI over sea areas and (ii) for computational efficiency, as sea gridboxes will be skipped before calculations. #' #' The landmask must be a grid spatially consistent with the input multigrid. You can use #' \code{\link[transformeR]{interpGrid}} in combination with the \code{getGrid} method to ensure this condition is fulfilled. . Its \code{data} component can be either a 2D or 3D array with the \code{dimensions} #' attribute \code{c("lat","lon")} or \code{c("time","lat","lon")} respectively. In the latter case, the length of the time #' dimension should be 1. Note that values of 0 correspond to sea areas (thus discarded for FWI calculation), being land areas any other #' values different from 0 (tipically 1 or 100, corresponding to the land/sea area fraction). #' #' \strong{Latitudinal chunking} #' #' Splitting the calculation in latitudinal chunks is highly advisable, and absolutely necessary when #' considering large spatial domains, otherwise running out of memory during the computation. The number #' of latitudinal chunks need to be estimated on a case-by-case basis, but in general there are no restrictions in the #' number of chunks that can be used, as long as it does not exceed the number of actual latitudes #' in the model grid. #' #' @template templateParallel #' #' @references #' \itemize{ #' \item Lawson, B.D. & Armitage, O.B., 2008. Weather guide for the Canadian Forest Fire Danger Rating System. Northern Forestry Centre, Edmonton (Canada). #' #' \item van Wagner, C.E., 1987. Development and structure of the Canadian Forest Fire Weather Index (Forestry Tech. Rep. No. 35). Canadian Forestry Service, Ottawa, Canada. #' #' \item van Wagner, C.E., Pickett, T.L., 1985. Equations and FORTRAN program for the Canadian forest fire weather index system (Forestry Tech. Rep. No. 33). Canadian Forestry Service, Ottawa, Canada. #' } #' #' @author J. Bedia \& M. Iturbide #' #' @export #' #' @importFrom abind abind asub #' @importFrom parallel parLapply splitIndices #' @importFrom transformeR redim getDim getShape parallelCheck getYearsAsINDEX subsetGrid array3Dto2Dmat mat2Dto3Darray fwiGrid <- function(multigrid, mask = NULL, what = "FWI", nlat.chunks = NULL, restart.annual = TRUE, parallel = FALSE, ncores = NULL, max.ncores = 16, ...) { what <- match.arg(what, choices = c("FFMC", "DMC", "DC", "ISI", "BUI", "FWI", "DSR"), several.ok = FALSE) fwi1D.opt.args <- list(...) months <- as.integer(substr(multigrid$Dates[[1]]$start, start = 6, stop = 7)) fwi1D.opt.args <- c(fwi1D.opt.args, list("what" = what)) if ("lat" %in% names(fwi1D.opt.args)) { message("NOTE: argument 'lat' will be overriden by the actual latitude of gridboxes\n(See help of fwi1D for details).") fwi1D.opt.args[-grep("lat", names(names(fwi1D.opt.args)))] } varnames <- multigrid$Variable$varName ycoords <- multigrid$xyCoords$y xcoords <- multigrid$xyCoords$x co <- expand.grid(ycoords, xcoords)[2:1] dimNames.mg <- getDim(multigrid) n.mem <- tryCatch(getShape(multigrid, "member"), error = function(er) 1L) ## if (n.mem == 1L) multigrid <- redim(multigrid) yrsindex <- getYearsAsINDEX(multigrid) nyears <- length(unique(yrsindex)) if (!is.null(mask)) { dimNames.mask <- getDim(mask) } if (is.null(nlat.chunks)) { nlat.chunks <- 1L } if (nlat.chunks <= 0) { nlat.chunks <- 1L message("Invalid 'nlat.chunks' argument value. It was ignored") } idx.chunk.list <- parallel::splitIndices(length(ycoords), nlat.chunks) if (any(vapply(idx.chunk.list, FUN = "length", FUN.VALUE = numeric(1)) < 2L)) { stop("Too many latitudinal chunks. Reduce the value of 'nlat.chunks' to a maximum of ", length(ycoords) %/% 2) } message("[", Sys.time(), "] Calculating ", what) aux.list <- lapply(1:nlat.chunks, function(k) { ## Lat chunking ind.lat <- idx.chunk.list[[k]] dims <- grep("lat", dimNames.mg) multigrid_chunk <- multigrid mask_chunk <- mask if (nlat.chunks > 1) { aux <- asub(multigrid$Data, idx = ind.lat, dims = dims) attr(aux, "dimensions") <- dimNames.mg multigrid_chunk$Data <- aux multigrid_chunk$xyCoords$y <- multigrid$xyCoords$y[ind.lat] ## Mask chunking if (!is.null(mask)) { aux <- asub(mask$Data, idx = ind.lat, dims = grep("lat", dimNames.mask)) attr(aux, "dimensions") <- dimNames.mask mask_chunk$Data <- aux mask_chunk$xyCoords$y <- mask_chunk$xyCoords$y[ind.lat] } aux <- NULL } ## Multigrid subsetting Tm1 <- subsetGrid(multigrid_chunk, var = grep("tas", varnames, value = TRUE)) Tm1 <- redim(Tm1, drop = FALSE) H1 <- subsetGrid(multigrid_chunk, var = grep("hurs", varnames, value = TRUE)) H1 <- redim(H1, drop = FALSE) r1 <- subsetGrid(multigrid_chunk, var = "tp") r1 <- redim(r1, drop = FALSE) W1 <- subsetGrid(multigrid_chunk, var = "wss") W1 <- redim(W1, drop = FALSE) multigrid_chunk <- NULL ## Parallel checks parallel.pars <- parallelCheck(parallel, max.ncores, ncores) if (n.mem < 2 && isTRUE(parallel.pars$hasparallel)) { parallel.pars$hasparallel <- FALSE message("NOTE: parallel computing only applies to multimember grids. The option was ignored") } if (parallel.pars$hasparallel) { apply_fun <- function(...) { parallel::parLapply(cl = parallel.pars$cl, ...) } on.exit(parallel::stopCluster(parallel.pars$cl)) } else { apply_fun <- lapply } ## Landmask if (!is.null(mask)) { if (!("^time" %in% dimNames.mask)) { aux <- unname(abind(mask_chunk$Data, along = 0L)) attr(aux, "dimensions") <- c("time", dimNames.mask) } else { aux <- mask_chunk$Data } msk <- array3Dto2Dmat(aux)[1,] ind <- which(msk > 0) msk <- NULL } else { aux <- suppressWarnings(subsetGrid(Tm1, members = 1))$Data aux <- array3Dto2Dmat(aux) ind <- which(apply(aux, MARGIN = 2, FUN = function(y) !all(is.na(y)))) } aux <- NULL ## FWI calculation message("[", Sys.time(), "] Processing chunk ", k, " out of ", nlat.chunks, "...") a <- apply_fun(1:n.mem, function(x) { Tm2 <- array3Dto2Dmat(subsetGrid(Tm1, members = x)$Data) H2 <- array3Dto2Dmat(subsetGrid(H1, members = x)$Data) r2 <- array3Dto2Dmat(subsetGrid(r1, members = x)$Data) W2 <- array3Dto2Dmat(subsetGrid(W1, members = x)$Data) b <- array(dim = dim(Tm2)) if (length(ind) > 0) { for (i in 1:length(ind)) { if (isTRUE(restart.annual)) { ## Iterate over years annual.list <- lapply(1:nyears, function(j) { idx <- which(yrsindex == unique(yrsindex)[j]) arg.list2 <- list("months" = months[idx], "Tm" = Tm2[idx,ind[i]], "H" = H2[idx,ind[i]], "r" = r2[idx,ind[i]], "W" = W2[idx,ind[i]], "lat" = co[ind[i],2]) arg.list <- c(fwi1D.opt.args, arg.list2) z <- tryCatch({suppressWarnings(drop(do.call("fwi1D", args = arg.list)))}, error = function(err) {rep(NA, length(idx))}) ## if (length(z) < length(idx)) z <- rep(NA, length(idx)) return(z) }) b[,ind[i]] <- do.call("c", annual.list) } else { arg.list2 <- list("months" = months, "Tm" = Tm2[,ind[i]], "H" = H2[,ind[i]], "r" = r2[,ind[i]], "W" = W2[,ind[i]], "lat" = co[ind[i],2]) arg.list <- c(fwi1D.opt.args, arg.list2) z <- tryCatch({suppressWarnings(drop(do.call("fwi1D", args = arg.list)))}, error = function(err) {rep(NA, length(months))}) ## if (length(z) < nrow(b)) z <- rep(NA, nrow(b)) b[,ind[i]] <- z } } out <- mat2Dto3Darray(mat2D = b, x = Tm1$xyCoords$x, y = Tm1$xyCoords$y) return(out) } }) Tm1 <- r1 <- H1 <- W1 <- NULL unname(do.call("abind", list(a, along = 0))) }) message("[", Sys.time(), "] Done.") ## Final grid and metadata fwigrid <- redim(subsetGrid(multigrid, var = varnames[1]), drop = FALSE) multigrid <- NULL dimNames <- getDim(fwigrid) fwigrid$Data <- unname(do.call("abind", c(aux.list, along = grep("lat", dimNames)))) aux.list <- NULL attr(fwigrid$Data, "dimensions") <- dimNames fwigrid$Variable <- list() fwigrid$Variable$varName <- what fwigrid$Variable$level <- NA desc <- switch(what, "FFMC" = "Fine Fuel Moisture Code", "DMC" = "Duff Moisture Code", "DC" = "Drought Code", "ISI" = "Initial Spread Index", "BUI" = "Builtup Index", "FWI" = "Fire Weather Index", "DSR" = "Daily Severity Rating") attr(fwigrid$Variable, "use_dictionary") <- FALSE attr(fwigrid$Variable, "description") <- desc attr(fwigrid$Variable, "units") <- "adimensional" attr(fwigrid$Variable, "longname") <- paste(desc, "component of the Canadian Fire Weather Index System") attr(fwigrid, "calculation") <- "Calculated with the fireDanger package (https://github.com/SantanderMetGroup/fireDanger)" return(fwigrid) }
aa50d822c149a368b699b8800e38a7fe889ba0bd
b75ec7c515a9fa9f7f988577d20e602e9c839240
/letters-learn-then-test/modeling/run-models.R
7f61e05ebef6b6a5031014134a4e5a5da9b2988a
[]
no_license
jodeleeuw/dynamic-constraints
1c6af747551945db71c1fe9fa83681cc687b2ea4
ffbde719a8df4e71713b3189851748f0748f1720
refs/heads/master
2021-01-21T04:47:19.711750
2016-06-30T18:31:42
2016-06-30T18:31:42
49,590,213
0
0
null
null
null
null
UTF-8
R
false
false
5,146
r
run-models.R
#### package requirements #### require(V8) # to run JS code require(jsonlite) # to parse JSON output from JavaScript require(plyr) # data storage manipulation #### load script that can generate sequences #### source('modeling/sequence-generators.R') #### set parameters for model testing #### reps_per_condition <- 1000 reps_per_item_in_seq <- 25 # behavioral experiment was 25 # PARSER parameters PARSER_maximum_percept_size <- 3 PARSER_initial_lexicon_weight <- 1 PARSER_shaping_weight_threshold <- 1 PARSER_reinforcement_rate <- 0.5 PARSER_forgetting_rate <- 0.05 PARSER_interference_rate <- 0.005 PARSER_logging <- "false" # MDLChunker parameters MDL_perceptual_span <- 25 MDL_memory_span <- 150 MDL_logging <- "false" # TRACX parameters #### functions to run each model once #### run_PARSER <- function(model, seq, condition) { if(condition=='seeded'){ seed_str <- ",[{word:['D','E','F'], weight: 100.0},{word:['G','H','I'], weight: 100.0},{word:['J','K','L'], weight: 100.0}]" } else { seed_str <- "" } ct <- new_context(); ct$source(model) ct$eval(paste0("PARSER.setup('",seq,"',{", "maximum_percept_size:",PARSER_maximum_percept_size,",", "initial_lexicon_weight:",PARSER_initial_lexicon_weight,",", "shaping_weight_threshold:",PARSER_shaping_weight_threshold,",", "reinforcement_rate:",PARSER_reinforcement_rate,",", "forgetting_rate:",PARSER_forgetting_rate,",", "interference_rate:",PARSER_interference_rate,",", "logging:",PARSER_logging, "}",seed_str,");")) ct$eval("PARSER.run()") #lexicon <- fromJSON(ct$eval("JSON.stringify(PARSER.getLexicon())")) base_weight <- as.numeric(ct$eval("PARSER.getWordStrength('A')"))+as.numeric(ct$eval("PARSER.getWordStrength('B')"))+as.numeric(ct$eval("PARSER.getWordStrength('C')")) target_weight <- as.numeric(ct$eval("PARSER.getWordStrength('ABC')")) + base_weight foil_weight <- base_weight # filter weights below 1.0 if(target_weight < 1.0){ target_weight <- 0 } if(foil_weight < 1.0){ foil_weight <- 0 } return(list(model="PARSER", condition=condition, target=target_weight, foil=foil_weight)) } run_MDLChunker <- function(model, seq, condition) { if(condition=='seeded'){ seq <- paste0(seed_sequence(100), seq) } ct <- new_context(); ct$source(model) ct$eval(paste0("MDLChunker.setup('",seq,"',{", "memory_span:",MDL_memory_span,",", "perceptual_span:",MDL_perceptual_span,",", "logging:",MDL_logging, "});")) ct$eval("MDLChunker.run()") #lexicon <- fromJSON(ct$eval("JSON.stringify(MDLChunker.getLexicon())")) target_weight <- as.numeric(ct$eval("MDLChunker.getCodeLengthForString('ABC')")) foil_weight <- as.numeric(ct$eval("MDLChunker.getCodeLengthForString('ACB')")) + as.numeric(ct$eval("MDLChunker.getCodeLengthForString('BAC')")) + as.numeric(ct$eval("MDLChunker.getCodeLengthForString('BCA')")) + as.numeric(ct$eval("MDLChunker.getCodeLengthForString('CBA')")) + as.numeric(ct$eval("MDLChunker.getCodeLengthForString('CAB')")) foil_weight <- foil_weight / 5 #memory <- fromJSON(ct$eval("MDLChunker.getMemory()")) #print(memory) return(list(model="MDLChunker",condition=condition,target=target_weight, foil=foil_weight)) } # run_TRACX <- function(model, seq, condition) { # # # if(condition=='seeded'){ # # seq.prepend <- # # } # # ct <- new_context(); # ct$source('modeling/models/TRACX-dependencies/sylvester.js') # ct$source('modeling/models/TRACX-dependencies/seedrandom-min.js') # ct$source(model) # ct$eval(paste0("TRACX.setTrainingData('",seq,"');")) # ct$eval('TRACX.getInputEncodings();') # ct$eval('TRACX.setTestData({Words:"ABC", PartWords: "", NonWords: "DHL"})') # ct$eval('TRACX.setSingleParameter("randomSeed","")') # ct$eval('TRACX.reset()') # lexicon <- fromJSON(ct$eval('JSON.stringify(TRACX.runFullSimulation(function(i,m){}))')) # target_weight <- as.numeric(lexicon$Words$mean) # foil_weight <- as.numeric(lexicon$NonWords$mean) # return(list(model="TRACX",condition=condition,target=target_weight, foil=foil_weight)) # } #### run all models #### # vectors to store data runs <- list() length(runs) <- reps_per_condition * 2 * 2 # 1 = number of models, 2 = number of conditions counter <- 1 # run models for(i in 1:reps_per_condition){ cat(paste('\r',i,'of',reps_per_condition)) four_seq <- four_triple_sequence(reps_per_item_in_seq) runs[[counter]] <- run_PARSER('modeling/models/parser.js',four_seq, 'unseeded') counter <- counter + 1 runs[[counter]] <- run_PARSER('modeling/models/parser.js',four_seq, 'seeded') counter <- counter + 1 runs[[counter]] <- run_MDLChunker('modeling/models/mdlchunker.js', four_seq, 'unseeded') counter <- counter + 1 runs[[counter]] <- run_MDLChunker('modeling/models/mdlchunker.js', four_seq, 'seeded') counter <- counter + 1 } model_run_data <- ldply(runs, data.frame) save(model_run_data, file='modeling/output/model_run_data.Rdata')
c18a5ab9e8b20890259a1ade1ff84d21974e6edc
71abde1c9025f7ab6d13f074192101dbab41c32d
/R Codes/r-tutorial-src/rtutor-pt1-c02.R
f1dada8b7131e368ee1c983eabfd19d7795c9496
[ "MIT" ]
permissive
hejibo/Psychological-Statistics
844ce22f8b70a860ba033b279b2f8f1823459062
2e245228f0e9d599ffaa50d01f41e3cdfbd3b17a
refs/heads/master
2021-05-09T10:14:53.475249
2018-05-01T01:27:00
2018-05-01T01:27:00
118,957,025
1
1
null
null
null
null
UTF-8
R
false
false
1,240
r
rtutor-pt1-c02.R
########################################################### # # Copyright (C) 2012 by Chi Yau # All rights reserved # # http://www.r-tutor.com # ################################ # c02 c(2, 3, 5) c(TRUE, FALSE, TRUE, FALSE, FALSE) c("aa", "bb", "cc", "dd", "ee") length(c("aa", "bb", "cc", "dd", "ee")) ################################ # c02-s01 n = c(2, 3, 5) s = c("aa", "bb", "cc", "dd", "ee") c(n, s) ################################ # c02-s02 a = c(1, 3, 5, 7) b = c(1, 2, 4, 8) 5 * a a + b a - b a * b a / b u = c(10, 20, 30) v = c(1, 2, 3, 4, 5, 6, 7, 8, 9) u + v w = c(10, 20, 30, 40) w + v ################################ # c02-s03 s = c("aa", "bb", "cc", "dd", "ee") s[3] s[-3] s[10] ################################ # c02-s04 s = c("aa", "bb", "cc", "dd", "ee") s[c(2, 3)] s[c(2, 3, 3)] s[c(2, 1, 3)] s[2:4] ################################ # c02-s05 s = c("aa", "bb", "cc", "dd", "ee") L = c(FALSE, TRUE, FALSE, TRUE, FALSE) s[L] s[c(FALSE, TRUE, FALSE, TRUE, FALSE)] ################################ # c02-s06 v = c("Mary", "Sue") v names(v) = c("First", "Last") v v["First"] v[c("Last", "First")]
3bbeb7f598690ebc091e06b29e2831a11b725aff
dd04fe1a24ad591212149ce4bd00f12c19a589db
/R/generics.R
f977f0c5d264e41b3f66f46929521dc8ad393565
[]
no_license
akgold/ynabR
87d6876b3daca41f6621df12dde222dbcdf4ee85
5c1f28df57293bf54184f079d8b52e0dcc1f92c4
refs/heads/master
2020-04-12T15:39:36.074593
2018-12-23T22:08:18
2018-12-23T22:08:18
162,588,489
1
0
null
null
null
null
UTF-8
R
false
false
215
r
generics.R
set_token <- function(x, token) { UseMethod("set_token") } get_token <- function(x) { UseMethod("get_token") } download <- function(x) { UseMethod("download") } parse <- function(x) { UseMethod("parse") }
7ee2f3b83fee5c2f435edc3f9cc9d94d838587a2
fe9d04e42bf96f0b490d1a0f5e94889b20e8420d
/R/load.NOAA.OISST.V2.R
d5424643882f01770aa4fff944b56805e51ed510
[]
no_license
mdsumner/probGLS
b8172865a6a497849bbb7a1d3edcda2a0245c815
b3808f6037dedacb48afa6c4155c87c06aa5bc2c
refs/heads/master
2020-12-31T22:11:07.263901
2018-11-29T07:43:37
2018-11-29T07:43:37
null
0
0
null
null
null
null
UTF-8
R
false
false
6,839
r
load.NOAA.OISST.V2.R
#' load NOAA OISST V2 #' #' This function takes 1-year-long NetCDF files from the #' ftp://ftp.cdc.noaa.gov/Datasets/noaa.oisst.v2.highres/ directory #' @param fname full path to NetCDF data file #' @param lsmask full path to land-sea mask NetCDF file #' @param lonW western-most longitude of search area, must be smaller than lonE #' @param lonE eastern-most longitude of search area, must be larger than lonW #' @param latS southern-most latitude of search area, must be smaller than latN #' @param latN northern-most latitude of search area, must be larger than latS #' @param date1 first date in file to extract, must be Date class #' @param date2 last date in file to extract, must be Date class #' @param use.landmask use land mask TRUE or FALSE #' @param extract.value which data to extract: "sst" - SST, "err" - SST error, "icec" - sea ice concentration #' @return A 3-dimensional array with latitudes in rows, longitudes in columns, and dates along the 3rd dimension. The value [1,1,1] is the northernmost, westernmost lat/long location on the 1st date. The value [1,1,2] is the 2nd date at the same lat/long location (if more than 1 date is requested). #' @return To extract lat/lon/date values from the output array, use the dimnames() function: #' @return lats = as.numeric(dimnames(sst2)$Lat) #' @return lons = as.numeric(dimnames(sst2)$Long) #' @return dates = as.Date(dimnames(sst2)$Date) #' @return #' @return NetCDF files should be downloaded from the links on: #' @return http://www.esrl.noaa.gov/psd/data/gridded/data.noaa.oisst.v2.highres.html #' @return In addition to the temperature data files, also download a copy of the landmask file lsmask.oisst.v2.nc from the same page. Inside the NetCDF files, data are available on a 0.25 degree latitude x 0.25 degree longitude global grid (720x1440 cells) From -89.875N to 89.875N, 0.125E to 359.875E. Locations are at the CENTER of a grid cell. #' @return modified after Luke Miller accessed Nov 25, 2014; https://github.com/millerlp/Misc_R_scripts/blob/master/NOAA_OISST_ncdf4.R #' @export load.NOAA.OISST.V2 = function(fname,lsmask,lonW,lonE,latS,latN, date1, date2,use.landmask=F, extract.value='sst'){ # Generate set of grid cell latitudes (center of cell) from south to north lats = seq(-89.875,89.875,0.25) # Generate set of grid cell longitudes (center of cell) lons = seq(0.125,359.875,0.25) # Create connection to NetCDF data file nc = nc_open(fname) lonWindx = which.min(abs(lonW - lons)) #get index of nearest longitude value lonEindx = which.min(abs(lonE - lons)) # Get index of nearest longitude value to lonE latSindx = which.min(abs(latS - lats)) #get index of nearest latitude value latNindx = which.min(abs(latN - lats)) # Get index of nearest latitude value to latN # The lon/lat indx values should now correspond to indices in the NetCDF # file for the desired grid cell. nlon = (lonEindx - lonWindx) + 1 # get number of longitudes to extract nlat = (latNindx - latSindx) + 1 # get number of latitudes to extract # Extract available dates from netCDF file ncdates = nc$dim$time$vals ncdates = as.Date(ncdates,origin = '1800-1-1') #available time points in nc if (class(date1) == 'Date'){ # Get index of nearest time point date1indx = which.min(abs(date1 - ncdates)) } else if (class(date1) == 'character'){ # Convert to a Date object first date1 = as.Date(date1) date1indx = which.min(abs(date1 - ncdates)) } if (missing(date2)) { # If date2 isn't specified, reuse date1 date2indx = which.min(abs(date1 - ncdates)) cat('Only 1 date specified\n') } else { if (class(date2) == 'Date'){ # If date2 exists, get index of nearest time point to date2 date2indx = which.min(abs(date2 - ncdates)) } else if (class(date2) == 'character'){ date2 = as.Date(date2) date2indx = which.min(abs(date2 - ncdates)) } } ndates = (date2indx - date1indx) + 1 #get number of time steps to extract # Define the output array sstout = array(data = NA, dim = c(nlon,nlat,ndates)) # Extract the data from the NetCDF file sstout[,,] = ncvar_get(nc, varid = extract.value, start = c(lonWindx,latSindx,date1indx), count = c(nlon,nlat,ndates)) # close SST ncdf nc_close(nc) # If there are missing data in the NetCDF, they should appear as 32767. # Replace that value with NA if it occurs anywhere. sstout = ifelse(sstout == 32767, NA, sstout) # Get dimensions of sstout array dims = dim(sstout) if(use.landmask==T) { nc2 = nc_open(lsmask) # Create array to hold land-sea mask mask = array(data = NA, dim = c(nlon,nlat,1)) # Get land-sea mask values (0 or 1) mask[,,] = ncvar_get(nc2, varid = "lsmask", start = c(lonWindx,latSindx,1), count = c(nlon,nlat,1)) #close land mask nc_close(nc2) # Replace 0's with NA's mask = ifelse(mask == 0,NA,1) for (i in 1:dims[3]) sstout[,,i] = sstout[,,i] * mask[,,1] # All masked values become NA } for (i in 1:dims[3]){ # Add dimension names attr(sstout,'dimnames') = list(Long = seq(lons[lonWindx],lons[lonEindx], by = 0.25), Lat = seq(lats[latSindx],lats[latNindx], by = 0.25), Date = as.character(seq(ncdates[date1indx], ncdates[date2indx],by = 1))) } # Rearrange the output matrix or array so that latitudes run from north to # south down the rows, and longitudes run from west to east across columns. # Make new output array to hold rearranged data. The dimension names will # match the newly rearranged latitude and longitude values sst2 = array(data = NA, dim = c(dims[2],dims[1],dims[3]), dimnames = list(Lat = rev(seq(lats[latSindx],lats[latNindx], by = 0.25)), Long = seq(lons[lonWindx],lons[lonEindx],by = 0.25), Date = as.character(seq(ncdates[date1indx], ncdates[date2indx],by = 1)))) # Step through each page of array and rearrange lat/lon values for (i in 1:dims[3]){ # Extract one day's worth of lat/lon pairs temp = as.matrix(sstout[,,i]) temp = t(temp) # transpose lon/lat to lat/lon temp = temp[nrow(temp):1,] # reverse row order to reverse latitudes sst2[,,i] = temp # write data to sst2 array } ########################## sst2 # return sst2 array ########################## } # end of function
f975c1c9b36013432baf837ff37caa63ea9f64af
48cf04931d5a0adbae40837f1322a689722e50a9
/man/pcg.Rd
36a3c0e53ea5698257e74d916d0e9d80725e83d7
[ "MIT" ]
permissive
natbprice/pcg2
93c3e9160261ac652943562516f48ab68a6bf108
15e884c008a83d024fb9fac449aaf366c8c5fc08
refs/heads/master
2020-05-04T18:37:58.024373
2019-04-03T20:05:40
2019-04-03T20:05:40
179,360,915
0
0
null
null
null
null
UTF-8
R
false
true
776
rd
pcg.Rd
% Generated by roxygen2: do not edit by hand % Please edit documentation in R/pcg.R \name{pcg} \alias{pcg} \title{Preconditioned conjugate gradient method solver} \usage{ pcg(Ax, b, M, x0, maxiter = 1000, tol = 1e-06) } \arguments{ \item{Ax}{function that takes argument x and returns matrix product A*x} \item{b}{right hand side of the linear system} \item{M}{preconditioner for A} \item{x0}{starting guess for solution} \item{maxiter}{maximum number of iterations} \item{tol}{tolerance for convergence on norm(residual, "2")} } \description{ Solve system of linear equations $Ax = b$ } \examples{ A <- matrix(c(4, 1, 1, 3), nrow = 2) b <- c(1, 2) x0 <- c(2, 1) Ax <- function(x) { A \%*\% x } M <- matrix(c(4, 0, 0, 3), nrow = nrow(A)) opt <- pcg(Ax, b, M, x0) }
37a5f94dc1b8a275654d2179647e7d33e0d9da92
8f9d93dc6c224cfc6f49cf82c7c650c80581032b
/PA_curves.R
7b2082f3de00b7dbb61b09e72a429367788c7201
[]
no_license
JackieOuzman/book_chap_weeds
57655b2e5cbb2c462b67ff1e7a75a6d79583efb5
9a7d40e609c5fc854e685d97eddca4e6b1d03d76
refs/heads/master
2020-05-05T12:27:35.865941
2019-04-09T05:19:11
2019-04-09T05:19:11
180,029,719
0
0
null
null
null
null
UTF-8
R
false
false
32,100
r
PA_curves.R
library(tidyverse) library(dplyr) library(ggplot2) #install.packages('DT') library(DT) library(readxl) #test #setwd("W:/Weeds/book_chapter_data") setwd("C:/Users/ouz001/book_chapter_data") #postcodes_with_coord <- read.csv("../Australian_Post_Codes_Lat_Lon.csv") #postcodes_with_coord_dist <- distinct(postcodes_with_coord, postcode, .keep_all = TRUE) #write.csv(postcodes_with_coord_dist, "postcodes_with_coord_dist.csv") #bring in the PA study advisors_PA <- read.csv("C:/Users/ouz001/book_chapter_data/PA/PA_advisors_adoption.csv") advisors_PA <- select(advisors_PA, farmers = Respondents, GRDC_RegionQ3, state, region_ReZoned, X12postcodes) glimpse(advisors_PA) #Try bring in more data from orginal study PA_survey <- read_excel("C:/Users/ouz001/book_chapter_data/PA/PA_data_rawish.xlsx") #SELECT JUST A FEW VARAIABLES FROM SURVEY PA_survey <- select(PA_survey, farmers = Respondents, Agro_Yr_StartQ34, Yr_No_Till_PA = NoTill_YrQ20, Yr_AutoSteer = Asteer_YrQ46, Yr_yld_map = Ymap_StartYearQ54, Yr_soil_testing = SoilTest_StartYearQ75) #JOIN THESE TWO TOGETHER glimpse(advisors_PA) glimpse(PA_survey) PA_survey = left_join(advisors_PA, PA_survey, by = "farmers") glimpse(PA_survey) #NOW GET THE ZONE AND REGIONS - so I need a file that has farmer XYadvisors_postcodes_join <- read.csv("C:/Users/ouz001/book_chapter_data/adoption_data/XYadvisors_postcodes_join.csv") %>% select(farmers = farmer, postcode, study, AGROECOLOG, REGIONS, state ) glimpse(XYadvisors_postcodes_join) ##JOIN TO THE PA SURVEY #### glimpse(XYadvisors_postcodes_join) glimpse(PA_survey) PA_survey_zone = left_join(XYadvisors_postcodes_join, PA_survey, by = "farmers") %>% select(farmers, postcode, study, state = state.x, AGROECOLOG, REGIONS, Agro_Yr_StartQ34, Yr_No_Till_PA, Yr_AutoSteer, Yr_yld_map, Yr_soil_testing) glimpse(PA_survey_zone) PA_survey_zone <- mutate(PA_survey_zone, Yr_Agro =as.factor(Agro_Yr_StartQ34), Yr_No_Till_PA=as.factor(Yr_No_Till_PA), Yr_AutoSteer_PA=as.factor(Yr_AutoSteer), Yr_yld_map_PA=as.factor(Yr_yld_map), Yr_soil_testing_PA=as.factor(Yr_soil_testing)) ###I THINK THIS IS THE DATA SET I NEED TO USE##### ####PA_survey_zone##### ###NOW I WANT TO CHECK OUT HOW THIS THESE adoption curves compare### #use data clm called PA_survey_zone # use zone as states #number of farmers in state # year_as_factor is the adoption clm I want to use #glimpse(PA_survey_zone) #test <- count(PA_survey_zone, Yr_Agro) fun_test2 <- function(df, zone, numb_farmers) { df_subset <- filter(df, state == zone) count_adoption_year <- count(df_subset,"Yr_Agro") count_adoption_year <- select(count_adoption_year, year = Yr_Agro, freq) #this clm year is a factor years_of_study <- data.frame(year = 1950:2014, #this is int id = 1:65, zone = zone) years_of_study <- mutate(years_of_study, year = as.factor(year)) adoption_df <- left_join(years_of_study, count_adoption_year, by= "year" ) adoption_df1 <- mutate(adoption_df, freq = replace_na(adoption_df$freq, 0)) adoption_df2 <- mutate(adoption_df1, cummulative = cumsum(adoption_df1$freq)) adoption_df3 <- mutate(adoption_df2, cumm_percent = (adoption_df2$cummulative/numb_farmers)*100) } ###PA Advisors BY state###### ###PA data for agro use - chcek to see if it looks the same as other work - yes so far so good. PA_AgroNSW <- fun_test2(PA_survey_zone, "NSW", 105) PA_AgroSA <- fun_test2(PA_survey_zone, "SA", 186) PA_AgroVIC <- fun_test2(PA_survey_zone, "VIC", 150) PA_AgroWA <- fun_test2(PA_survey_zone, "WA", 128) PA_Agro_state <- rbind(PA_AgroNSW, PA_AgroSA, PA_AgroVIC, PA_AgroWA) %>% mutate(year = as.integer(year)) glimpse(PA_Agro_state) ggplot(PA_Agro_state, aes(year, cumm_percent))+ geom_line()+ theme_classic()+ theme(legend.position = "bottom")+ xlim(1980, 2015)+ ylim(0,100)+ facet_wrap(.~zone)+ labs(x = "Years", y = "percenatge of farmers", title = "Adoption of advisor use per states", subtitle = "This study has no farmers in QLD") ###PA Yr_No_Till_PA BY state###### fun_test3 <- function(df, zone, numb_farmers) { df_subset <- filter(df, state == zone) count_adoption_year <- count(df_subset,"Yr_No_Till_PA") count_adoption_year <- select(count_adoption_year, year = Yr_No_Till_PA, freq) #this clm year is a factor years_of_study <- data.frame(year = 1950:2014, #this is int id = 1:65, zone = zone) years_of_study <- mutate(years_of_study, year = as.factor(year)) adoption_df <- left_join(years_of_study, count_adoption_year, by= "year" ) adoption_df1 <- mutate(adoption_df, freq = replace_na(adoption_df$freq, 0)) adoption_df2 <- mutate(adoption_df1, cummulative = cumsum(adoption_df1$freq)) adoption_df3 <- mutate(adoption_df2, cumm_percent = (adoption_df2$cummulative/numb_farmers)*100) } Yr_No_Till_PANSW <- fun_test3(PA_survey_zone, "NSW", 105) Yr_No_Till_PASA <- fun_test3(PA_survey_zone, "SA", 186) Yr_No_Till_PAVIC <- fun_test3(PA_survey_zone, "VIC", 150) Yr_No_Till_PAWA <- fun_test3(PA_survey_zone, "WA", 128) Yr_No_Till_PA_state <- rbind(Yr_No_Till_PANSW, Yr_No_Till_PASA, Yr_No_Till_PAVIC, Yr_No_Till_PAWA) %>% mutate(year = as.integer(year)) glimpse(Yr_No_Till_PA_state) ggplot(Yr_No_Till_PA_state, aes(year, cumm_percent))+ geom_line()+ theme_classic()+ theme(legend.position = "bottom")+ xlim(1980, 2015)+ ylim(0,100)+ facet_wrap(.~zone) ###put the no till data from PA survey over the advisor PA data### glimpse(PA_Agro_state) glimpse(Yr_No_Till_PA_state) PA_Agro_state <- mutate(PA_Agro_state, adoption = "advisors") Yr_No_Till_PA_state <- mutate(Yr_No_Till_PA_state, adoption = "No_till_PA") PA_no_till_advisors <- rbind(PA_Agro_state, Yr_No_Till_PA_state ) ggplot(PA_no_till_advisors, aes(year, cumm_percent, group = adoption))+ geom_line(aes(linetype=adoption))+ theme_classic()+ theme(legend.position = "bottom")+ xlim(1980, 2015)+ ylim(0,100)+ facet_wrap(.~zone) ##this says its the same as the no till study in 2014 ###OK for the 2014 weeds adoption data### no till XYNoTill_postcodes_join <- read.csv("C:/Users/ouz001/book_chapter_data/adoption_data/XYNoTill_postcodes_join_GRDC_SLA.csv") #change the year from interger to factor XYNoTill_postcodes_join <- mutate(XYNoTill_postcodes_join, year_as_factor = as.factor(Q15_year_f)) glimpse(XYNoTill_postcodes_join) adoption_curve_function_NoTill_S <- function(df, zone, numb_farmers) { df_subset <- filter(df, state == zone) count_adoption_year <- count(df_subset,"year_as_factor") count_adoption_year <- select(count_adoption_year, year = year_as_factor, freq) years_of_study <- data.frame(year = 1950:2014, id = 1:65, zone = zone) years_of_study <- mutate(years_of_study, year = as.factor(year)) adoption_df <- left_join(years_of_study, count_adoption_year, by= "year" ) adoption_df <- mutate(adoption_df, freq = replace_na(adoption_df$freq, 0)) adoption_df <- mutate(adoption_df, cummulative = cumsum(adoption_df$freq)) adoption_df <- mutate(adoption_df, cumm_percent = (adoption_df$cummulative/numb_farmers)*100) adoption_df$year <- as.double(adoption_df$year) #return(df_subset) return(adoption_df) } QLD_NT_weeds <- adoption_curve_function_NoTill_S(XYNoTill_postcodes_join, "QLD", 59) NSW_NT_weeds <- adoption_curve_function_NoTill_S(XYNoTill_postcodes_join, "NSW", 153) SA_NT_weeds <- adoption_curve_function_NoTill_S(XYNoTill_postcodes_join, "SA", 65) VIC_NT_weeds <- adoption_curve_function_NoTill_S(XYNoTill_postcodes_join, "VIC", 141) WA_NT_weeds <- adoption_curve_function_NoTill_S(XYNoTill_postcodes_join, "WA", 179) NoTill_adoption_state_NT_weeds <- rbind(QLD_NT_weeds, NSW_NT_weeds, SA_NT_weeds,VIC_NT_weeds, WA_NT_weeds ) NoTill_adoption_state_NT_weeds <- mutate(NoTill_adoption_state_NT_weeds, adoption = "No_till_Weeds") ####Join adoption data for weeds - no till , PA no till and advisors glimpse(PA_no_till_advisors) glimpse(NoTill_adoption_state_NT_weeds) PA_no_till_advisors_weeds_noTill <- rbind(PA_no_till_advisors, NoTill_adoption_state_NT_weeds) ggplot(PA_no_till_advisors_weeds_noTill, aes(year, cumm_percent, group = adoption))+ geom_line(aes(linetype=adoption))+ theme_classic()+ theme(legend.position = "bottom")+ xlim(1980, 2015)+ ylim(0,100)+ facet_wrap(.~zone) ggsave("1Adoption_no_till_and_advisor_PA_No_till_weeds.png", width = 9.8, height = 5.6, units = "in") glimpse(PA_survey_zone) ###PA Autosteer Yr_AutoSteer_PA fun_test4 <- function(df, zone, numb_farmers) { df_subset <- filter(df, state == zone) count_adoption_year <- count(df_subset,"Yr_AutoSteer_PA") count_adoption_year <- select(count_adoption_year, year = Yr_AutoSteer_PA, freq) #this clm year is a factor years_of_study <- data.frame(year = 1950:2014, #this is int id = 1:65, zone = zone) years_of_study <- mutate(years_of_study, year = as.factor(year)) adoption_df <- left_join(years_of_study, count_adoption_year, by= "year" ) adoption_df1 <- mutate(adoption_df, freq = replace_na(adoption_df$freq, 0)) adoption_df2 <- mutate(adoption_df1, cummulative = cumsum(adoption_df1$freq)) adoption_df3 <- mutate(adoption_df2, cumm_percent = (adoption_df2$cummulative/numb_farmers)*100) } ###PA Autosteer Yr_AutoSteer_PANSW <- fun_test4(PA_survey_zone, "NSW", 105) Yr_AutoSteer_PASA <- fun_test4(PA_survey_zone, "SA", 186) Yr_AutoSteer_PAVIC <- fun_test4(PA_survey_zone, "VIC", 150) Yr_AutoSteer_PAWA <- fun_test4(PA_survey_zone, "WA", 128) Yr_AutoSteer_PA_state <- rbind(Yr_AutoSteer_PANSW, Yr_AutoSteer_PASA, Yr_AutoSteer_PAVIC, Yr_AutoSteer_PAWA) %>% mutate(year = as.integer(year)) glimpse(Yr_No_Till_PA_state) Yr_AutoSteer_PA_state <- mutate(Yr_AutoSteer_PA_state, adoption = "PA_Auto_steer") glimpse(Yr_AutoSteer_PA_state) glimpse(NoTill_adoption_state_NT_weeds) Yr_AutoSteer_PA_weeds_noTill <- rbind(Yr_AutoSteer_PA_state, NoTill_adoption_state_NT_weeds, PA_Agro_state) ggplot(Yr_AutoSteer_PA_weeds_noTill, aes(year, cumm_percent, group = adoption))+ geom_line(aes(linetype=adoption))+ theme_classic()+ theme(legend.position = "bottom")+ xlim(1980, 2015)+ ylim(0,100)+ facet_wrap(.~zone) ggsave("Yr_AutoSteer_PA_weeds_noTill.png", width = 9.8, height = 5.6, units = "in") Yr_AutoSteer_PA_Agro <- rbind(Yr_AutoSteer_PA_state, PA_Agro_state) ggplot(Yr_AutoSteer_PA_Agro, aes(year, cumm_percent, group = adoption))+ geom_line(aes(linetype=adoption))+ theme_classic()+ theme(legend.position = "bottom")+ xlim(1980, 2015)+ ylim(0,100)+ facet_wrap(.~zone) ggsave("Yr_AutoSteer_PA_Agro.png", width = 9.8, height = 5.6, units = "in") ###Yr_yld_map glimpse(PA_survey_zone) fun_test5 <- function(df, zone, numb_farmers) { df_subset <- filter(df, state == zone) count_adoption_year <- count(df_subset,"Yr_yld_map_PA") count_adoption_year <- select(count_adoption_year, year = Yr_yld_map_PA, freq) #this clm year is a factor years_of_study <- data.frame(year = 1950:2014, #this is int id = 1:65, zone = zone) years_of_study <- mutate(years_of_study, year = as.factor(year)) adoption_df <- left_join(years_of_study, count_adoption_year, by= "year" ) adoption_df1 <- mutate(adoption_df, freq = replace_na(adoption_df$freq, 0)) adoption_df2 <- mutate(adoption_df1, cummulative = cumsum(adoption_df1$freq)) adoption_df3 <- mutate(adoption_df2, cumm_percent = (adoption_df2$cummulative/numb_farmers)*100) } ###PA Yr_yld_map Yr_yld_map_PANSW <- fun_test5(PA_survey_zone, "NSW", 105) Yr_yld_map_PASA <- fun_test5(PA_survey_zone, "SA", 186) Yr_yld_map_PAVIC <- fun_test5(PA_survey_zone, "VIC", 150) Yr_yld_map_PAWA <- fun_test5(PA_survey_zone, "WA", 128) Yr_yld_map_PA_state <- rbind(Yr_yld_map_PANSW, Yr_yld_map_PASA, Yr_yld_map_PAVIC, Yr_yld_map_PAWA) %>% mutate(year = as.integer(year)) glimpse(Yr_yld_map_PA_state) Yr_yld_map_PA_state <- mutate(Yr_yld_map_PA_state, adoption = "PA_Yld_map") glimpse(Yr_yld_map_PA_state) glimpse(NoTill_adoption_state_NT_weeds) Yr_yld_map_PA_weeds_noTill <- rbind(Yr_yld_map_PA_state, NoTill_adoption_state_NT_weeds) ggplot(Yr_yld_map_PA_weeds_noTill, aes(year, cumm_percent, group = adoption))+ geom_line(aes(linetype=adoption))+ theme_classic()+ theme(legend.position = "bottom")+ xlim(1980, 2015)+ ylim(0,100)+ facet_wrap(.~zone) ggsave("Yr_yld_map_PA_weeds_noTill.png", width = 9.8, height = 5.6, units = "in") Yr_yld_map_PA_Agro_state <- rbind(Yr_yld_map_PA_state, PA_Agro_state) ggplot(Yr_yld_map_PA_Agro_state, aes(year, cumm_percent, group = adoption))+ geom_line(aes(linetype=adoption))+ theme_classic()+ theme(legend.position = "bottom")+ xlim(1980, 2015)+ ylim(0,100)+ facet_wrap(.~zone) ggsave("Yr_yld_map_PA_Agro_state.png", width = 9.8, height = 5.6, units = "in") ###Yr_soil_testing_PA glimpse(PA_survey_zone) fun_test6 <- function(df, zone, numb_farmers) { df_subset <- filter(df, state == zone) count_adoption_year <- count(df_subset,"Yr_soil_testing_PA") count_adoption_year <- select(count_adoption_year, year = Yr_soil_testing_PA, freq) #this clm year is a factor years_of_study <- data.frame(year = 1950:2014, #this is int id = 1:65, zone = zone) years_of_study <- mutate(years_of_study, year = as.factor(year)) adoption_df <- left_join(years_of_study, count_adoption_year, by= "year" ) adoption_df1 <- mutate(adoption_df, freq = replace_na(adoption_df$freq, 0)) adoption_df2 <- mutate(adoption_df1, cummulative = cumsum(adoption_df1$freq)) adoption_df3 <- mutate(adoption_df2, cumm_percent = (adoption_df2$cummulative/numb_farmers)*100) } ###PA Yr_soil_testing_PA Yr_soil_testing_PANSW <- fun_test6(PA_survey_zone, "NSW", 105) Yr_soil_testing_PASA <- fun_test6(PA_survey_zone, "SA", 186) Yr_soil_testing_PAVIC <- fun_test6(PA_survey_zone, "VIC", 150) Yr_soil_testing_PAWA <- fun_test6(PA_survey_zone, "WA", 128) Yr_soil_testing_PA_state <- rbind(Yr_soil_testing_PANSW, Yr_soil_testing_PASA, Yr_soil_testing_PAVIC, Yr_soil_testing_PAWA) %>% mutate(year = as.integer(year)) glimpse(Yr_soil_testing_PA_state) Yr_soil_testing_PA_state <- mutate(Yr_soil_testing_PA_state, adoption = "PA_soil_test") glimpse(Yr_soil_testing_PA_state) glimpse(NoTill_adoption_state_NT_weeds) Yr_soil_testing_PA_state_weeds_noTill <- rbind(Yr_soil_testing_PA_state, NoTill_adoption_state_NT_weeds) ggplot(Yr_soil_testing_PA_state_weeds_noTill, aes(year, cumm_percent, group = adoption))+ geom_line(aes(linetype=adoption))+ theme_classic()+ theme(legend.position = "bottom")+ xlim(1980, 2015)+ ylim(0,100)+ facet_wrap(.~zone) ggsave("Yr_soil_testing_PA_weeds_noTill.png", width = 9.8, height = 5.6, units = "in") Yr_soil_testing_PA_Agro_state <- rbind(Yr_soil_testing_PA_state, PA_Agro_state) ggplot(Yr_soil_testing_PA_Agro_state, aes(year, cumm_percent, group = adoption))+ geom_line(aes(linetype=adoption))+ theme_classic()+ theme(legend.position = "bottom")+ xlim(1980, 2015)+ ylim(0,100)+ facet_wrap(.~zone) ggsave("Yr_soil_testing_PA_Agro_state.png", width = 9.8, height = 5.6, units = "in") ###OK for the 2014 weeds adoption data### no till XYNoTill_postcodes_join <- read.csv("C:/Users/ouz001/book_chapter_data/adoption_data/XYNoTill_postcodes_join_GRDC_SLA.csv") %>% select(farmer, postcode, study, AGROECOLOG, REGIONS, state) glimpse(XYNoTill_postcodes_join) crop_top <- read_excel("C:/Users/ouz001/book_chapter_data/Weeds/Raw_data_Weeds_with_postcodes.xlsx")%>% select(KEY, Q20l1) crop_top <- mutate(crop_top, Yr_crop_top = if_else(Q20l1 ==-99, 0, Q20l1)) glimpse(crop_top) crop_top <- mutate(crop_top, Yr_crop_top = as.factor(Yr_crop_top), farmer = KEY) #join the data togther to get clm for crop topping and for states etc.. crop_top1 <- left_join(crop_top, XYNoTill_postcodes_join, by = "farmer") glimpse(crop_top1) adoption_curve_fun7 <- function(df, zone, numb_farmers) { df_subset <- filter(df, state == zone) count_adoption_year <- count(df_subset,"Yr_crop_top") count_adoption_year <- select(count_adoption_year, year = Yr_crop_top, freq) years_of_study <- data.frame(year = 1950:2014, id = 1:65, zone = zone) years_of_study <- mutate(years_of_study, year = as.factor(year)) adoption_df <- left_join(years_of_study, count_adoption_year, by= "year" ) adoption_df <- mutate(adoption_df, freq = replace_na(adoption_df$freq, 0)) adoption_df <- mutate(adoption_df, cummulative = cumsum(adoption_df$freq)) adoption_df <- mutate(adoption_df, cumm_percent = (adoption_df$cummulative/numb_farmers)*100) adoption_df$year <- as.double(adoption_df$year) #return(df_subset) return(adoption_df) } QLD_croptop <- adoption_curve_fun7(crop_top1, "QLD", 59) NSW_croptop <- adoption_curve_fun7(crop_top1, "NSW", 153) SA_croptop <- adoption_curve_fun7(crop_top1, "SA", 65) VIC_croptop <- adoption_curve_fun7(crop_top1, "VIC", 141) WA_croptop <- adoption_curve_fun7(crop_top1, "WA", 179) croptop_states <- rbind(QLD_croptop, NSW_croptop, SA_croptop,VIC_croptop, WA_croptop ) croptop_states <- mutate(croptop_states, adoption = "crop top") ggplot(croptop_states, aes(year, cumm_percent))+ geom_line(aes(linetype=adoption))+ #scale_linetype_manual(values=c("dashed", "dotted", "solid"))+ theme_classic()+ theme(legend.position = "bottom")+ xlim(1980, 2015)+ ylim(0,100)+ facet_wrap(.~zone)+ labs(x = "Years", y = "Percentage of farmers") glimpse(NoTill_adoption_state_NT_weeds) croptop_advisors <- rbind(PA_Agro_state, croptop_states) croptop_advisors <- filter(croptop_advisors, zone != "QLD") glimpse(croptop_advisors) ggplot(croptop_advisors, aes(year, cumm_percent, group = adoption))+ geom_line(aes(linetype=adoption))+ #scale_linetype_manual(values=c("dashed", "dotted", "solid"))+ theme_classic()+ theme(legend.position = "bottom")+ xlim(1980, 2015)+ ylim(0,100)+ facet_wrap(.~zone)+ labs(x = "Years", y = "Percentage of farmers") ggsave("croptop_advisors.png", width = 9.8, height = 5.6, units = "in") ####make a data set that has Advisor (PA study), no till, soil test, auto steer and crop topping###### glimpse(PA_Agro_state) glimpse(NoTill_adoption_state_NT_weeds) glimpse(Yr_soil_testing_PA_state) glimpse(Yr_AutoSteer_PA_state) glimpse(croptop_states) Agro_noTill_soil_test_autoSteer_crop_top <- rbind(PA_Agro_state, NoTill_adoption_state_NT_weeds, Yr_soil_testing_PA_state, Yr_AutoSteer_PA_state, croptop_states) Agro_noTill_soil_test_autoSteer_crop_top<- filter(Agro_noTill_soil_test_autoSteer_crop_top, zone!= "QLD") ggplot(Agro_noTill_soil_test_autoSteer_crop_top, aes(year, cumm_percent, group = adoption))+ geom_line(aes(linetype=adoption))+ #scale_linetype_manual(values=c("dashed", "dotted", "solid"))+ theme_classic()+ theme(legend.position = "bottom")+ xlim(1980, 2015)+ ylim(0,100)+ facet_wrap(.~zone)+ labs(x = "Years", y = "Percentage of farmers") ggsave("Agro_noTill_soil_test_autoSteer_crop_top.png", width = 9.8, height = 5.6, units = "in") #########make a data set that has Advisor (PA study), no till, soil test, and auto steer ###### glimpse(PA_Agro_state) glimpse(NoTill_adoption_state_NT_weeds) glimpse(Yr_soil_testing_PA_state) glimpse(Yr_AutoSteer_PA_state) Agro_noTill_soil_test_autoSteer <- rbind(PA_Agro_state, NoTill_adoption_state_NT_weeds, Yr_soil_testing_PA_state, Yr_AutoSteer_PA_state) Agro_noTill_soil_test_autoSteer<- filter(Agro_noTill_soil_test_autoSteer, zone!= "QLD") ggplot(Agro_noTill_soil_test_autoSteer, aes(year, cumm_percent, group = adoption))+ geom_line(aes(linetype=adoption))+ #scale_linetype_manual(values=c("dashed", "dotted", "solid"))+ theme_classic()+ theme(legend.position = "bottom")+ xlim(1980, 2015)+ ylim(0,100)+ facet_wrap(.~zone)+ labs(x = "Years", y = "Percentage of farmers") ggsave("Agro_noTill_soil_test_autoSteer.png", width = 9.8, height = 5.6, units = "in") ####### Narrow windrow buring ########### XYNoTill_postcodes_join <- read.csv("C:/Users/ouz001/book_chapter_data/adoption_data/XYNoTill_postcodes_join_GRDC_SLA.csv") %>% select(farmer, postcode, study, AGROECOLOG, REGIONS, state) glimpse(XYNoTill_postcodes_join) narrow_windrow_burn <- read_excel("C:/Users/ouz001/book_chapter_data/Weeds/Raw_data_Weeds_with_postcodes.xlsx")%>% select(KEY, Q21c1 ) narrow_windrow_burn <- mutate(narrow_windrow_burn, Yr_narrow_windrow_burn = if_else(Q21c1 ==-99, 0, Q21c1)) glimpse(narrow_windrow_burn) narrow_windrow_burn <- mutate(narrow_windrow_burn, Yr_narrow_windrow_burn = as.factor(Yr_narrow_windrow_burn), farmer = KEY) #join the data togther to get clm for narrow windrow and for states etc.. narrow_windrow_burn1 <- left_join(narrow_windrow_burn, XYNoTill_postcodes_join, by = "farmer") glimpse(narrow_windrow_burn1) adoption_curve_fun7 <- function(df, zone, numb_farmers) { df_subset <- filter(df, state == zone) count_adoption_year <- count(df_subset,"Yr_narrow_windrow_burn") count_adoption_year <- select(count_adoption_year, year = Yr_narrow_windrow_burn, freq) years_of_study <- data.frame(year = 1950:2014, id = 1:65, zone = zone) years_of_study <- mutate(years_of_study, year = as.factor(year)) adoption_df <- left_join(years_of_study, count_adoption_year, by= "year" ) adoption_df <- mutate(adoption_df, freq = replace_na(adoption_df$freq, 0)) adoption_df <- mutate(adoption_df, cummulative = cumsum(adoption_df$freq)) adoption_df <- mutate(adoption_df, cumm_percent = (adoption_df$cummulative/numb_farmers)*100) adoption_df$year <- as.double(adoption_df$year) #return(df_subset) return(adoption_df) } QLD_narrow_burn <- adoption_curve_fun7(narrow_windrow_burn1, "QLD", 59) NSW_narrow_burn <- adoption_curve_fun7(narrow_windrow_burn1, "NSW", 153) SA_narrow_burn <- adoption_curve_fun7(narrow_windrow_burn1, "SA", 65) VIC_narrow_burn <- adoption_curve_fun7(narrow_windrow_burn1, "VIC", 141) WA_narrow_burn <- adoption_curve_fun7(narrow_windrow_burn1, "WA", 179) narrow_burn_states <- rbind(QLD_narrow_burn, NSW_narrow_burn, SA_narrow_burn,VIC_narrow_burn, WA_narrow_burn ) narrow_burn_states <- mutate(narrow_burn_states, adoption = "narrow windrow burning") ####### Narrow windrow buring, crop top and advisor(PA) by states ####### glimpse(PA_Agro_state) glimpse(narrow_burn_states) Agro_advisor_crop_top_narrow_burn <- rbind(PA_Agro_state, narrow_burn_states, croptop_states) Agro_advisor_crop_top_narrow_burn <- filter(Agro_advisor_crop_top_narrow_burn, zone!= "QLD") ggplot(Agro_advisor_crop_top_narrow_burn , aes(year, cumm_percent, group = adoption))+ geom_line(aes(linetype=adoption))+ #scale_linetype_manual(values=c("solid", "dashed", "dotted" ))+ theme_classic()+ theme(legend.position = "bottom")+ theme(legend.position = "")+ xlim(1980, 2015)+ ylim(0,100)+ facet_wrap(.~zone)+ labs(x = "Years", y = "Percentage of farmers") ggsave("xxAgro_advisor_crop_top_narrow_burn.png", width = 9.8, height = 5.6, units = "in") glimpse(Agro_advisor_crop_top_narrow_burn) Agro_advisor_crop_top_narrow_burn_not_NSW_VIC <- filter(Agro_advisor_crop_top_narrow_burn, zone == "SA" | zone == "WA") glimpse(Agro_advisor_crop_top_narrow_burn_not_NSW_VIC) ggplot(Agro_advisor_crop_top_narrow_burn_not_NSW_VIC , aes(year, cumm_percent, group = adoption))+ geom_line(aes(linetype=adoption))+ #scale_linetype_manual(values=c("solid", "dashed", "dotted" ))+ theme_classic()+ theme(legend.position = "bottom")+ theme(legend.position = "")+ xlim(1980, 2015)+ ylim(0,100)+ facet_wrap(.~zone)+ labs(x = "Years", y = "Percentage of farmers") ggsave("xxAgro_advisor_crop_top_narrow_burn_SA_WA.png", width = 9.8, height = 5.6, units = "in") ########### Burn advisors and crop top by regions ######## ###BURN XYNoTill_postcodes_join <- read.csv("C:/Users/ouz001/book_chapter_data/adoption_data/XYNoTill_postcodes_join_GRDC_SLA.csv") %>% select(farmer, postcode, study, AGROECOLOG, REGIONS, state) glimpse(XYNoTill_postcodes_join) narrow_windrow_burn <- read_excel("C:/Users/ouz001/book_chapter_data/Weeds/Raw_data_Weeds_with_postcodes.xlsx")%>% select(KEY, Q21c1 ) narrow_windrow_burn <- mutate(narrow_windrow_burn, Yr_narrow_windrow_burn = if_else(Q21c1 ==-99, 0, Q21c1)) glimpse(narrow_windrow_burn) narrow_windrow_burn <- mutate(narrow_windrow_burn, Yr_narrow_windrow_burn = as.factor(Yr_narrow_windrow_burn), farmer = KEY) #join the data togther to get clm for narrow windrow and for states etc.. narrow_windrow_burn1 <- left_join(narrow_windrow_burn, XYNoTill_postcodes_join, by = "farmer") glimpse(narrow_windrow_burn1) count(narrow_windrow_burn1,"REGIONS") adoption_curve_fun7 <- function(df, zone, numb_farmers) { df_subset <- filter(df, REGIONS == zone) count_adoption_year <- count(df_subset,"Yr_narrow_windrow_burn") count_adoption_year <- select(count_adoption_year, year = Yr_narrow_windrow_burn, freq) years_of_study <- data.frame(year = 1950:2014, id = 1:65, zone = zone) years_of_study <- mutate(years_of_study, year = as.factor(year)) adoption_df <- left_join(years_of_study, count_adoption_year, by= "year" ) adoption_df <- mutate(adoption_df, freq = replace_na(adoption_df$freq, 0)) adoption_df <- mutate(adoption_df, cummulative = cumsum(adoption_df$freq)) adoption_df <- mutate(adoption_df, cumm_percent = (adoption_df$cummulative/numb_farmers)*100) adoption_df$year <- as.double(adoption_df$year) #return(df_subset) return(adoption_df) } glimpse(narrow_windrow_burn1) Northern_narrow_burn <- adoption_curve_fun7(narrow_windrow_burn1, "Northern", 118) Southern_narrow_burn <- adoption_curve_fun7(narrow_windrow_burn1, "Southern", 298) Western_narrow_burn <- adoption_curve_fun7(narrow_windrow_burn1, "Western", 172) narrow_burn_Regions <- rbind(Northern_narrow_burn, Southern_narrow_burn, Western_narrow_burn ) narrow_burn_Regions <- mutate(narrow_burn_Regions, adoption = "narrow windrow burning") ###crop XYNoTill_postcodes_join <- read.csv("C:/Users/ouz001/book_chapter_data/adoption_data/XYNoTill_postcodes_join_GRDC_SLA.csv") %>% select(farmer, postcode, study, AGROECOLOG, REGIONS, state) glimpse(XYNoTill_postcodes_join) crop_top <- read_excel("C:/Users/ouz001/book_chapter_data/Weeds/Raw_data_Weeds_with_postcodes.xlsx")%>% select(KEY, Q20l1) crop_top <- mutate(crop_top, Yr_crop_top = if_else(Q20l1 ==-99, 0, Q20l1)) glimpse(crop_top) crop_top <- mutate(crop_top, Yr_crop_top = as.factor(Yr_crop_top), farmer = KEY) #join the data togther to get clm for crop topping and for states etc.. crop_top1 <- left_join(crop_top, XYNoTill_postcodes_join, by = "farmer") glimpse(crop_top1) adoption_curve_fun7 <- function(df, zone, numb_farmers) { df_subset <- filter(df, REGIONS == zone) count_adoption_year <- count(df_subset,"Yr_crop_top") count_adoption_year <- select(count_adoption_year, year = Yr_crop_top, freq) years_of_study <- data.frame(year = 1950:2014, id = 1:65, zone = zone) years_of_study <- mutate(years_of_study, year = as.factor(year)) adoption_df <- left_join(years_of_study, count_adoption_year, by= "year" ) adoption_df <- mutate(adoption_df, freq = replace_na(adoption_df$freq, 0)) adoption_df <- mutate(adoption_df, cummulative = cumsum(adoption_df$freq)) adoption_df <- mutate(adoption_df, cumm_percent = (adoption_df$cummulative/numb_farmers)*100) adoption_df$year <- as.double(adoption_df$year) #return(df_subset) return(adoption_df) } Northern_croptop <- adoption_curve_fun7(crop_top1, "Northern", 118) Southern_croptop <- adoption_curve_fun7(crop_top1, "Southern", 298) Western_croptop <- adoption_curve_fun7(crop_top1, "Western", 172) croptop_Region <- rbind(Northern_croptop, Southern_croptop, Western_croptop ) croptop_Region <- mutate(croptop_Region, adoption = "crop top") #### ADVISOR glimpse(PA_survey_zone) count(PA_survey_zone,"REGIONS") fun_test2 <- function(df, zone, numb_farmers) { df_subset <- filter(df, REGIONS == zone) count_adoption_year <- count(df_subset,"Yr_Agro") count_adoption_year <- select(count_adoption_year, year = Yr_Agro, freq) #this clm year is a factor years_of_study <- data.frame(year = 1950:2014, #this is int id = 1:65, zone = zone) years_of_study <- mutate(years_of_study, year = as.factor(year)) adoption_df <- left_join(years_of_study, count_adoption_year, by= "year" ) adoption_df1 <- mutate(adoption_df, freq = replace_na(adoption_df$freq, 0)) adoption_df2 <- mutate(adoption_df1, cummulative = cumsum(adoption_df1$freq)) adoption_df3 <- mutate(adoption_df2, cumm_percent = (adoption_df2$cummulative/numb_farmers)*100) } ###PA Advisors BY Regions###### Southern_advisor <- fun_test2(PA_survey_zone, "Southern", 441) Western_advisor <- fun_test2(PA_survey_zone, "Western", 128) advisor_region <- rbind(Southern_advisor, Western_advisor ) advisor_region <- mutate(advisor_region, adoption = "advisor", year = as.double(year)) glimpse(advisor_region) glimpse(narrow_burn_Regions) glimpse(croptop_Region) glimpse(advisor_region) advisor_narrow_crop_region <- rbind(advisor_region,narrow_burn_Regions, croptop_Region ) glimpse(advisor_narrow_crop_region) advisor_narrow_crop_region_St_West <- filter(advisor_narrow_crop_region,zone != "Northern" ) glimpse(advisor_narrow_crop_region_St_West) ggplot(advisor_narrow_crop_region_St_West , aes(year, cumm_percent, group = adoption))+ geom_line(aes(linetype=adoption))+ #scale_linetype_manual(values=c("solid", "dashed", "dotted" ))+ theme_classic()+ #theme(legend.position = "bottom")+ theme(legend.position = "")+ xlim(1980, 2015)+ ylim(0,100)+ facet_wrap(.~zone)+ labs(x = "Years", y = "Percentage of farmers") ggsave("xxAgro_advisor_crop_top_narrow_burn_Sth_West.png", width = 9.8, height = 5.6, units = "in")
e50d4809b08ed0aee617b9d7e8100e04c35b72df
a159106592c73eef0699c2485ce781ef092c6022
/04-limma-voom.R
574bcf311cb7161eb600e377eeb44773fbb3d8fa
[ "MIT" ]
permissive
bobia9991/RNA-seq
4a0f87be27fbff876899f393940c2500f21049a2
c29df866ba949a3e6abe118ef4c21f34162628e0
refs/heads/main
2023-06-19T20:16:32.091529
2021-07-16T00:38:11
2021-07-16T00:38:11
560,504,965
1
0
MIT
2022-11-01T16:42:15
2022-11-01T16:42:14
null
UTF-8
R
false
false
16,056
r
04-limma-voom.R
# RNA-seq pipeline # Ben Laufer # Modifies and expands on these references: #https://ucdavis-bioinformatics-training.github.io/2018-June-RNA-Seq-Workshop/thursday/DE.html #https://www.bioconductor.org/packages/devel/workflows/vignettes/RNAseq123/inst/doc/limmaWorkflow.html # Load packages ----------------------------------------------------------- setwd("~/Box/PEBBLES/RNA") packages <- c("edgeR", "tidyverse", "RColorBrewer", "org.Mm.eg.db", "AnnotationDbi", "EnhancedVolcano", "enrichR", "openxlsx", "gt", "glue", "Glimma", "sva", "DMRichR") enrichR:::.onAttach() # Needed or else "EnrichR website not responding" stopifnot(suppressMessages(sapply(packages, require, character.only = TRUE))) #BiocManager::install("ben-laufer/DMRichR") # To test and develop, assign the variable tissue and then just run the main sections sink("RNA-seq_log.txt", type = "output", append = FALSE, split = TRUE) tidyr::crossing(tissue = c("placenta", "brain"), sex = c("male", "female")) %>% purrr::pwalk(function(tissue, sex){ dir.create(glue::glue("{tissue}_{sex}")) # Count Matrix ------------------------------------------------------------ #name <- gsub( "(?:[^_]+_){4}([^_ ]+)*$","", files) # STAR quantMode geneCounts output: #column 1: gene ID #column 2: counts for unstranded RNA-seq #column 3: counts for the 1st read strand aligned with RNA (htseq-count option -s yes) #column 4: counts for the 2nd read strand aligned with RNA (htseq-count option -s reverse) # KAPA mRNA HyperPrep Kit reads are reverse stranded, so select column 4 # Confirm by looking at the N_noFeature line for the 3rd and 4th column and pick the column with the lowest count. sampleNames <- list.files(path = glue::glue(getwd(), "/GeneCounts"), pattern = "*.ReadsPerGene.out.tab") %>% stringr::str_split_fixed("_", n = 3) %>% tibble::as_tibble() %>% tidyr::unite(Name, c(V1:V2), sep = "-") %>% dplyr::select(Name) %>% purrr::flatten_chr() # Could alternatively use edgeR::readDGE() but that calls to the slower read.delim() ensemblIDs <- list.files(path = glue::glue(getwd(), "/GeneCounts"), pattern = "*.ReadsPerGene.out.tab", full.names = TRUE)[1] %>% data.table::fread(select = 1) %>% purrr::flatten_chr() countMatrix <- list.files(path = glue::glue(getwd(), "/GeneCounts"), pattern = "*.ReadsPerGene.out.tab", full.names = TRUE) %>% purrr::map_dfc(data.table::fread, select = 4, data.table = FALSE) %>% magrittr::set_colnames(sampleNames) %>% magrittr::set_rownames(ensemblIDs) # Remove meta info countMatrix <- countMatrix[-c(1:4),] # Design Matrix ----------------------------------------------------------- designMatrix <- readxl::read_xlsx("sample_info.xlsx") %>% dplyr::rename(group = Treatment) %>% dplyr::mutate_if(is.character, as.factor) %>% dplyr::mutate(Name = as.character(Name)) # # Recode sex # designMatrix$Sex <- as.character(designMatrix$Sex) # designMatrix$Sex[designMatrix$Sex == "F"] <- "0" # designMatrix$Sex[designMatrix$Sex == "M"] <- "1" # designMatrix$Sex <- as.factor(designMatrix$Sex) samples.idx <- pmatch(designMatrix$Name, colnames(countMatrix)) designMatrix <- designMatrix[order(samples.idx),] # Preprocessing ----------------------------------------------------------- print(glue::glue("Preprocessing {sex} {tissue} samples")) # Select sample subset designMatrix <- designMatrix %>% dplyr::filter(Tissue == tissue & Sex == sex) countMatrix <- countMatrix %>% dplyr::select(contains(designMatrix$Name)) %>% as.matrix # Create DGE list and calculate normalization factors countMatrix <- countMatrix %>% DGEList() %>% calcNormFactors() # Reorder design matrix samples.idx <- pmatch(designMatrix$Name, rownames(countMatrix$samples)) designMatrix <- designMatrix[order(samples.idx),] stopifnot(rownames(countMatrix$samples) == designMatrix$Name) # Add sample info from design matrix to DGE list countMatrix$samples <- countMatrix$samples %>% tibble::add_column(designMatrix %>% dplyr::select(-Name)) # Add gene info countMatrix$genes <- purrr::map_dfc(c("SYMBOL", "GENENAME", "ENTREZID", "CHR"), function(column){ rownames(countMatrix$counts) %>% AnnotationDbi::mapIds(org.Mm.eg.db, keys = ., column = column, keytype = 'ENSEMBL') %>% as.data.frame() %>% tibble::remove_rownames() %>% purrr::set_names(column) }) # Raw density of log-CPM values L <- mean(countMatrix$samples$lib.size) * 1e-6 M <- median(countMatrix$samples$lib.size) * 1e-6 logCPM <- cpm(countMatrix, log = TRUE) logCPM.cutoff <- log2(10/M + 2/L) nsamples <- ncol(countMatrix) col <- brewer.pal(nsamples, "Paired") pdf(glue::glue("{tissue}_{sex}/{tissue}_{sex}_density_plot.pdf"), height = 8.5, width = 11) par(mfrow = c(1,2)) plot(density(logCPM[,1]), col = col[1], lwd = 2, las = 2, main = "", xlab = "") title(main = "A. Raw data", xlab = "Log-cpm") abline(v = logCPM.cutoff, lty = 3) for (i in 2:nsamples){ den <- density(logCPM[,i]) lines(den$x, den$y, col = col[i], lwd = 2) } legend("topright", designMatrix$Name, text.col = col, bty = "n", cex = 0.5) # Filter genes with low expression rawCount <- dim(countMatrix) keep.exprs <- filterByExpr(countMatrix, group = countMatrix$samples$group, lib.size = countMatrix$samples$lib.size) countMatrix <- countMatrix[keep.exprs,, keep.lib.sizes = FALSE] %>% calcNormFactors() filterCount <- dim(countMatrix) print(glue::glue("{100 - round((filterCount[1]/rawCount[1])*100)}% of genes were filtered from {rawCount[2]} samples, \\ where there were {rawCount[1]} genes before filtering and {filterCount[1]} genes after filtering for {tissue}")) # Filtered density plot of log-CPM values logCPM <- cpm(countMatrix, log = TRUE) plot(density(logCPM[,1]), col = col[1], lwd = 2, las =2 , main = "", xlab = "") title(main = "B. Filtered data", xlab = "Log-cpm") abline(v = logCPM.cutoff, lty = 3) for (i in 2:nsamples){ den <- density(logCPM[,i]) lines(den$x, den$y, col = col[i], lwd = 2) } legend("topright", designMatrix$Name, text.col = col, bty = "n", cex = 0.5) dev.off() # Interactive MDS plot Glimma::glMDSPlot(countMatrix, groups = designMatrix, path = getwd(), folder = "interactivePlots", html = glue::glue("{tissue}_{sex}_MDS-Plot"), launch = FALSE) # Surrogate variables analysis -------------------------------------------- # # Create model matrices, with null model for svaseq, and don't force a zero intercept # mm <- model.matrix(~group + Litter, # data = designMatrix) # # mm0 <- model.matrix(~1 + Litter, # data = designMatrix) # # # svaseq requires normalized data that isn't log transformed # cpm <- cpm(countMatrix, log = FALSE) # # # Calculate number of surrogate variables # nSv <- num.sv(cpm, # mm, # method = "leek") # # # Estimate surrogate variables # svObj <- svaseq(cpm, # mm, # mm0, # n.sv = nSv) # # # Update model to include surrogate variables # mm <- model.matrix(~Treatment + svObj$sv, # data = designMatrix) # Voom transformation and calculation of variance weights ----------------- print(glue::glue("Normalizing {sex} {tissue} samples")) # Design mm <- model.matrix(~group, data = designMatrix) # Voom pdf(glue::glue("{tissue}_{sex}/{tissue}_{sex}_voom_mean-variance_trend.pdf"), height = 8.5, width = 11) voomLogCPM <- voom(countMatrix, mm, plot = TRUE) dev.off() # Make litter a random effect, since limma warns "coefficients not estimable" for some litters # Ref: https://support.bioconductor.org/p/11956/ # Obstacle: Cannot do this properly with surrogtate variables, since there's an error when including litter in null model # Duplicate correlations alternative for other scenarios: # https://support.bioconductor.org/p/68916/ # https://support.bioconductor.org/p/110987/ correlations <- duplicateCorrelation(voomLogCPM, mm, block = designMatrix$Litter) # Extract intraclass correlation within litters correlations <- correlations$consensus.correlation # Boxplots of logCPM values before and after normalization pdf(glue::glue("{tissue}_{sex}/{tissue}_{sex}_normalization_boxplots.pdf"), height = 8.5, width = 11) par(mfrow=c(1,2)) boxplot(logCPM, las = 2, col = col, main = "") title(main = "A. Unnormalised data", ylab = "Log-cpm") boxplot(voomLogCPM$E, las = 2, col = col, main = "") title(main = "B. Normalised data", ylab = "Log-cpm") dev.off() # Fitting linear models in limma ------------------------------------------ print(glue::glue("Testing {sex} {tissue} samples for differential expression")) # Weight standard errors of log fold changes by within litter correlation fit <- lmFit(voomLogCPM, mm, correlation = correlations, block = designMatrix$Litter) head(coef(fit)) # Save normalized expression values for WGCNA voomLogCPM$E %>% as.data.frame() %>% tibble::rownames_to_column(var = "Gene") %>% openxlsx::write.xlsx(glue::glue("{tissue}_{sex}/{tissue}_{sex}_voomLogCPMforWGCNA.xlsx")) # Create DEG tibble ------------------------------------------------------- print(glue::glue("Creating DEG list of {sex} {tissue} samples")) efit <- fit %>% contrasts.fit(coef = 2) %>% # Change for different models eBayes() # Final model plot pdf(glue::glue("{tissue}_{sex}/{tissue}_{sex}_final_model_mean-variance_trend.pdf"), height = 8.5, width = 11) plotSA(efit, main = "Final model: Mean-variance trend") dev.off() # Interactive MA plot Glimma::glimmaMA(efit, dge = countMatrix, path = getwd(), html = glue::glue("interactivePlots/{tissue}_{sex}_MDA-Plot.html"), launch = FALSE) # Top differentially expressed genes DEGs <- efit %>% topTable(sort.by = "P", n = Inf) %>% rownames_to_column() %>% tibble::as_tibble() %>% dplyr::rename(ensgene = rowname) %>% dplyr::mutate(FC = dplyr::case_when(logFC > 0 ~ 2^logFC, logFC < 0 ~ -1/(2^logFC))) %>% dplyr::select(SYMBOL, GENENAME, FC, logFC, P.Value, adj.P.Val, AveExpr, t, B, ensgene) %T>% openxlsx::write.xlsx(file = glue::glue("{tissue}_{sex}/{tissue}_{sex}_DEGs.xlsx")) # For a continuous trait the FC is the change per each unit # Volcano Plot ------------------------------------------------------------ volcano <- DEGs %>% EnhancedVolcano::EnhancedVolcano(title = "", labSize = 5, lab = .$SYMBOL, x = 'logFC', y = 'P.Value', # P.Value 'adj.P.Val' col = c("grey30", "royalblue", "royalblue", "red2"), pCutoff = 0.05, FCcutoff = 0.0) + ggplot2::coord_cartesian(xlim = c(-3, 3), ylim = c(0, 4)) ggplot2::ggsave(glue::glue("{tissue}_{sex}/{tissue}_{sex}_volcano.pdf"), plot = volcano, device = NULL, width = 11, height = 8.5) # HTML report ------------------------------------------------------------- print(glue::glue("Saving html report of {sex} {tissue} samples")) DEGs <- DEGs %>% dplyr::filter(P.Value < 0.05) %T>% openxlsx::write.xlsx(file = glue::glue("{tissue}_{sex}/{tissue}_{sex}_filtered_DEGs.xlsx")) DEGs %>% dplyr::rename(Gene = SYMBOL, "p-value" = P.Value, "adjusted p-value" = adj.P.Val, Description = GENENAME, ensembl = ensgene) %>% gt() %>% tab_header( title = glue::glue("{nrow(DEGs)} Differentially Expressed Genes"), subtitle = glue::glue("{round(sum(DEGs$logFC > 0) / nrow(DEGs), digits = 2)*100}% up-regulated, \\ {round(sum(DEGs$logFC < 0) / nrow(DEGs), digits = 2)*100}% down-regulated")) %>% fmt_number( columns = vars("FC", "logFC", "AveExpr", "t", "B"), decimals = 2) %>% fmt_scientific( columns = vars("p-value", "adjusted p-value"), decimals = 2) %>% as_raw_html(inline_css = TRUE) %>% write(glue::glue("{tissue}_{sex}/{tissue}_{sex}_DEGs.html")) # Heatmap ----------------------------------------------------------------- print(glue::glue("Plotting heatmap of {sex} {tissue} samples")) voomLogCPM$E[which(rownames(voomLogCPM$E) %in% DEGs$ensgene),] %>% as.matrix() %>% pheatmap::pheatmap(., scale = "row", annotation_col = designMatrix %>% tibble::column_to_rownames(var = "Name") %>% dplyr::select(Treatment = group, Litter), color = RColorBrewer::brewer.pal(11, name = "RdBu") %>% rev(), show_colnames = FALSE, show_rownames = F, #angle_col = 45, border_color = "grey", main = glue::glue("Z-Scores of {nrow(DEGs)} Differentially Expressed Genes"), fontsize = 16, filename = glue::glue("{tissue}_{sex}/{tissue}_{sex}_heatmap.pdf"), width = 11, height = 8.5, annotation_colors = list(Treatment = c("PCB" = "#F8766D", "Control" = "#619CFF"))) # Ontologies and Pathways ------------------------------------------------- print(glue::glue("Performing GO and pathway analysis of {sex} {tissue} samples")) tryCatch({ DEGs %>% dplyr::select(SYMBOL) %>% purrr::flatten() %>% enrichR::enrichr(c("GO_Biological_Process_2018", "GO_Cellular_Component_2018", "GO_Molecular_Function_2018", "KEGG_2019_Mouse", "Panther_2016", "Reactome_2016", "RNA-Seq_Disease_Gene_and_Drug_Signatures_from_GEO")) %T>% # %>% #purrr::map(~ dplyr::filter(., Adjusted.P.value < 0.05)) %>% #purrr::map(~ dplyr::filter(., stringr::str_detect(Genes, ";"))) %>% openxlsx::write.xlsx(file = glue::glue("{tissue}_{sex}/{tissue}_{sex}_enrichr.xlsx")) %>% DMRichR::slimGO(tool = "enrichR", annoDb = "org.Mm.eg.db", plots = FALSE) %T>% openxlsx::write.xlsx(file = glue::glue("{tissue}_{sex}/{tissue}_{sex}_rrvgo_enrichr.xlsx")) %>% DMRichR::GOplot() %>% ggplot2::ggsave(glue::glue("{tissue}_{sex}/{tissue}_{sex}_enrichr_plot.pdf"), plot = ., device = NULL, height = 8.5, width = 10) }, error = function(error_condition) { print(glue::glue("ERROR: Gene Ontology pipe didn't finish for {sex} {tissue}")) }) print(glue::glue("The pipeline has finished for {sex} {tissue} samples")) }) sink()
ac607c1ae2ef2c2d97078b0b0d1da6f8f392c87e
ee54a85e446285fcc568caffe2dab934089e12e2
/xp11/03_lme4_models_xp11.R
bcc00dd5f53b200f904f6b81238479bdcf90c1ef
[ "CC0-1.0" ]
permissive
bricebeffara/rwa_evaluative_conditioning_data_analysis
e8697aee9cb7097c19ec9207a092d547658246e2
81c35c84148f9579b68624e32c0efc0dbad43e3c
refs/heads/main
2023-06-03T01:15:34.778559
2021-06-17T09:55:12
2021-06-17T09:55:12
375,408,897
0
0
null
null
null
null
UTF-8
R
false
false
6,006
r
03_lme4_models_xp11.R
# File name: brms_models_xp11.R # Online archive: gitlab # Authors: Brice Beffara & Amélie Bret # Tue Jul 03 14:24:51 2018 ------------------------------ # Contact: brice.beffara@slowpen.science amelie.bret@univ-grenoble-alpes.fr http://slowpen.science # # This R script was used to build and compute brms models # corresponding to the 10th experiment of Amelie Bret's doctoral work # # This R script defines and computes brms models # main effects, interaction effects, and simple slopes of interest # # 3 posependent variables of interest : # RWA (continuous, centered and scaled) # usvalence : positive (0.5) vs. negative (-0.5) # and warning : no warn (-0.5) vs. warn (0.5) # # and 1 ordinal dependent variables : # Ratings of Greebles from 1 (very negative) to 9 (very positive) # # This program is believed to be free of errors, but it comes with no guarantee! # The user bears all responsibility for interpreting the results. # # This preambule is largely inspired by John K. Kruschke's work at https://osf.io/wp2ry/ # ### To run this program, please do the following: ### 1. Install the general-purpose programming language R from ### http://www.r-project.org/ ### Install the version of R appropriate for your computer's operating ### system (Wposows, MacOS, or Linux). ### 2. Install the R editor, RStudio, from ### http://rstudio.org/ ### This editor is not necessary, but highly recommended. ### 3. After the above actions are accomplished, this program should ### run as-is in R. You may "source" it to run the whole thing at once, ### or, preferably, run lines consecutively from the beginning. ################################################################################ # Loading packages needed (and installing if necessary) for this part p_load(lme4, # main package for models htmlTable, # helps to extract results xtable, install = TRUE, gridExtra, sjstats, sjmisc, update = getOption("pac_update"), character.only = FALSE) # In case we want to save summaries col2keep <- c("Estimate", "l-95% CI", "u-95% CI") #------------------------------------------------------------------------------------ # We run our first model for fixed main and interaction effects #------------------------------------------------------------------------------------ # model warn_resp_lme4 <- lmer(response ~ usvalence * warn * RWAscore + (1|ppt) + (1|stim1), data = warn_df) # Save summary & confint model_gen_xp11_lme4 <- round(cbind(summary(warn_resp_lme4)$coefficients, confint(warn_resp_lme4)[c(4:11),]), 2) # export output png("tables/lme4/model_gen_xp11_lme4.png", height=480, width=720) p<-tableGrob(model_gen_xp11_lme4) grid.arrange(p) dev.off() #------------------------------------------------------------------------------------ # Then we run our second step models to decompose interactions effects # We look at the interaction between RWA and usvalence at each level of conditioning #------------------------------------------------------------------------------------ #------------- ### RWA * valence in the !!no warning!! condition #------------- # model warn_resp_nowa_lme4 <- lmer(response ~ usvalence * no_warn * RWAscore + (1|ppt) + (1|stim1), data = warn_df) # Save summary & confint model_nowa_xp11_lme4 <- round(cbind(summary(warn_resp_nowa_lme4)$coefficients, confint(warn_resp_nowa_lme4)[c(4:11),]), 2) # export output png("tables/lme4/model_nowa_xp11_lme4.png", height=480, width=720) p<-tableGrob(model_nowa_xp11_lme4) grid.arrange(p) dev.off() #------------- ### RWA * usvalence in the !!warning!! condition #------------- # model warn_resp_yewa_lme4 <- lmer(response ~ usvalence * ye_warn * RWAscore + (1|ppt) + (1|stim1), data = warn_df) # Save summary & confint model_yewa_xp11_lme4 <- round(cbind(summary(warn_resp_yewa_lme4)$coefficients, confint(warn_resp_yewa_lme4)[c(4:11),]), 2) # export output png("tables/lme4/model_yewa_xp11_lme4.png", height=480, width=720) p<-tableGrob(model_yewa_xp11_lme4) grid.arrange(p) dev.off() #------------------------------------------------------------------------------------ # Then we run our third step model to decompose the interaction # between RWA and usvalence in the no warning conditioning condition # The interaction slope between RWA and usvalence includes 0 in the warning condition #------------------------------------------------------------------------------------ ############################# ##################### no warning ############################# #------------- ### Simple slope of RWA in the !!negative!! valence & !!no warning!! condition #------------- #model warn_resp_nowa_neg_lme4 <- lmer(response ~ usvalence_neg * no_warn * RWAscore + (1|ppt) + (1|stim1), data = warn_df) # Save summary & confint model_nowa_neg_xp11_lme4 <- round(cbind(summary(warn_resp_nowa_neg_lme4)$coefficients, confint(warn_resp_nowa_neg_lme4)[c(4:11),]), 2) # export output png("tables/lme4/model_nowa_neg_xp11_lme4.png", height=480, width=720) p<-tableGrob(model_nowa_neg_xp11_lme4) grid.arrange(p) dev.off() #------------- ### Simple slope of RWA in the !!positive!! valence & !!no warning!! condition #------------- #model warn_resp_nowa_pos_lme4 <- lmer(response ~ usvalence_pos * no_warn * RWAscore + (1|ppt) + (1|stim1), data = warn_df) # Save summary & confint model_nowa_pos_xp11_lme4 <- round(cbind(summary(warn_resp_nowa_pos_lme4)$coefficients, confint(warn_resp_nowa_pos_lme4)[c(4:11),]), 2) # export output png("tables/lme4/model_nowa_pos_xp11_lme4.png", height=480, width=720) p<-tableGrob(model_nowa_pos_xp11_lme4) grid.arrange(p) dev.off()
c26bcf647efddad8c4dee524523fa5d8652b3d59
9aafde089eb3d8bba05aec912e61fbd9fb84bd49
/codeml_files/newick_trees_processed/3710_0/rinput.R
8dc09fe6a3d6abe05808bd3269dee7312c1710b4
[]
no_license
DaniBoo/cyanobacteria_project
6a816bb0ccf285842b61bfd3612c176f5877a1fb
be08ff723284b0c38f9c758d3e250c664bbfbf3b
refs/heads/master
2021-01-25T05:28:00.686474
2013-03-23T15:09:39
2013-03-23T15:09:39
null
0
0
null
null
null
null
UTF-8
R
false
false
135
r
rinput.R
library(ape) testtree <- read.tree("3710_0.txt") unrooted_tr <- unroot(testtree) write.tree(unrooted_tr, file="3710_0_unrooted.txt")
df93f7eb0a1738ba9026c6fe3d67f295743bf4a1
4dcbd7bde2c131cb0d3c96990eb20df6e9ea22ed
/man/sub.obj.Rd
24a5d3ea1079c59842b887720ed7a3145a06c48b
[]
no_license
LindaLuck/VoxR
e3428d0967a2d5b5b3470cdfe85df92d75e7dfb9
4ad74f91aa421881120fc387137861d57c92f790
refs/heads/master
2022-12-24T02:44:23.030784
2020-09-30T07:50:08
2020-09-30T07:50:08
null
0
0
null
null
null
null
UTF-8
R
false
true
242
rd
sub.obj.Rd
% Generated by roxygen2: do not edit by hand % Please edit documentation in R/subobj.R \name{sub.obj} \alias{sub.obj} \title{Deprecated function} \usage{ sub.obj(...) } \arguments{ \item{...}{parameters} } \description{ Deprecated function }
42472f2d46987657ab4c973e7cfdc5a3450e836d
e434c57f4ca421735aaae1046f67c916e6b91bd4
/man/captioner-package.Rd
45bf6b77145d5107d62e7a5812a30c08ec7a8ac1
[]
no_license
achetverikov/captioner
ac03bafe1334744cb2152bb93de1eb13ed084203
ba2b2ceae048d8deafdf77d25752dab8cce81633
refs/heads/master
2021-05-30T02:48:12.645510
2015-02-10T22:52:00
2015-02-10T22:52:00
null
0
0
null
null
null
null
UTF-8
R
false
false
353
rd
captioner-package.Rd
% Generated by roxygen2 (4.0.2): do not edit by hand \docType{package} \name{captioner-package} \alias{captioner-package} \title{captioner: A package for numbering figures and generating captions} \description{ Contains the function \code{\link{captioner}} for generating numbered captions. } \author{ Alathea D Letaw, \email{alathea@zoology.ubc.ca} }
a03d090f5041f16e66b5ad9ab6462396605118c9
7ae3533e4a28daa4f27f8a36ad71bb0a03cab159
/MNN_clust.R
bfba82f3231c4c572947990e0d3884109993812e
[ "Apache-2.0" ]
permissive
Bionett/merge_ds_norm
75dc51b92e6b718da7360e2c7b82e606457bcd52
ab8eb23983e2280d5fabf5449949a8e854f10316
refs/heads/master
2020-06-04T01:49:13.565069
2019-06-13T19:37:34
2019-06-13T19:37:34
191,821,847
0
1
null
null
null
null
UTF-8
R
false
false
5,291
r
MNN_clust.R
# unsupervised clustering methods # semiunsupervised method ## Matching Mutual Nearest Neighbors #BiocManager::install("scran") library(scran) library(data.table) library(dplyr) library(scater) file2esets <- function(esetfile, inputfolder, sampattern, metafile, batch) { eset <- fread(paste0(inputfolder, esetfile)) meta <- fread(paste0(inputfolder, metafile)) bs <- unique(meta[[batch]]) setNames( lapply(bs, function(b) { bsampls <- dplyr::filter(meta, batch == b)[["Sample_geo_accession"]] select(eset, matches(paste(c("Symbol", bsampls), collapse = "|"))) }), bs) } #sapply(file2esets(fs[1], inputfolder, sampattern, metafile, batch), dim) eset2obj <- function(eset, sampattern) { emat <- data.matrix(select(eset, matches(sampattern))) rownames(emat) <- eset$Symbol sce <- SingleCellExperiment(list(counts = emat, logcounts = emat)) sce } do_mnn_corr <- function(esetfiles, inputfolder, sampattern, metafile, batch, ncores) { #mclapply(esetfiles, function(esetfile) { lapply(esetfiles, function(esetfile) { if(esetfile %in% c("blood_sjia_pjia_UPC_eset_sjia_sex_difF_REM_deg.tsv", "blood_sjia_pjia_yugene_eset_sjia_sex_difF_REM_deg.tsv", "blood_sjia_pjia_zscore_eset_sjia_sex_difF_REM_deg.tsv")) { NULL } else { print(esetfile) # get merged eset file and split it in several tables based on the GSE accessions print("starting...") esets <- file2esets(esetfile, inputfolder, sampattern, metafile, batch) print("esetfile splitted...") # from tables to SingleCellExpression objects, deprecating the GSE58667 that contains just 4 samples print(names(esets)) esetobjs <- lapply(esets[c("GSE55319", "GSE58667", "GSE80060_2011-03", "GSE80060", "GSE17590")], function(e) eset2obj(e, sampattern)) # MNN batch correction set.seed(224033911) print("correcting...") mnn.out <- do.call(fastMNN, c(esetobjs, list(k = 10, d = 50, approximate = TRUE, auto.order=TRUE))) print("corrected...") # Assembling original datset omat <- Reduce(cbind, lapply(esetobjs, function(eo) logcounts(eo))) # writing MNN corrected matrix print("writing MNN corrected eset...") cor.exp <- tcrossprod(mnn.out$rotation, mnn.out$corrected) colnames(cor.exp) <- colnames(omat) cor.exp <- cbind(data.frame("Symbol" = rownames(omat)), cor.exp) write.table(cor.exp, paste0("../result/MNN_corrected_", gsub("\\.tsv", "", esetfile), ".tsv"), sep = "\t", quote = FALSE, row.names = FALSE) print("visualizing...") # visualize cluster differences among the original and corrected datasets through tSNE plots sce <- SingleCellExperiment(list(counts = omat, logcounts = omat)) reducedDim(sce, "MNN") <- mnn.out$corrected sce$Batch <- as.character(mnn.out$batch) # Including disease metadata msce <- data.frame("Sample_geo_accession" = rownames(sce@colData)) # Reference metadata meta <- read.delim(paste0(inputfolder, metafile)) rownames(meta) <- meta$Sample_geo_accession msce <- merge(msce, meta, by = "Sample_geo_accession") sce$Disease <- as.character(msce$Characteristics..disease.) sce$scandate <- as.character(msce$scandate_yymm) set.seed(100) # Using irlba to set up the t-SNE, for speed. # visualization of the original data osce <- runPCA(sce, ntop=Inf, method="irlba") osce <- runTSNE(osce, use_dimred="PCA") ot <- plotTSNE(osce, colour_by="Batch") + ggtitle("Original") dot <- plotTSNE(osce, colour_by="Disease") + ggtitle("Original") sdot <- plotTSNE(osce, colour_by="scandate") + ggtitle("scandate") # Visualizartion of the MNN transformed data set.seed(100) csce <- runTSNE(sce, use_dimred="MNN") ct <- plotTSNE(csce, colour_by="Batch") + ggtitle("Corrected") dct <- plotTSNE(csce, colour_by="Disease") + ggtitle("Corrected") sdct <- plotTSNE(csce, colour_by="scandate") + ggtitle("scandate") # Clustering # The aim is to use the SNN graph to perform clustering of cells # via community detection algorithms in the igraph package snn.gr <- buildSNNGraph(sce, use.dimred = "MNN") clusters <- igraph::cluster_walktrap(snn.gr) table(clusters$membership, sce$Batch) csce$Cluster <- factor(clusters$membership) # Ploting cluster on tSNE plot cls <- plotTSNE(csce, colour_by="Cluster") # Ploting batch, disease and ad-hoc clustering on the original and MNN transformed data pdf(paste0("../result/tSNE_MNN_corrected_", gsub("\\.tsv", "", esetfile), ".pdf"), width = 15, height = 15) multiplot(ot, ct, sdot, cls, dot, dct, sdct, cls, cols=2) dev.off() #}, mc.cores = ncores) } }) } # Inputs inputfolder <- "../data/" esetpattern <- "^blood_sjia_pjia_" sampattern <- "GSM" metafile <- "metadata_blood_sjia_pjia_normal_samples.tsv" batch <- "batch" #"GSE_accession_wfound" esetfiles <- list.files(path = inputfolder, pattern = esetpattern) ncores <- 3 do_mnn_corr(esetfiles[1:4], inputfolder, sampattern, metafile, batch, ncores)
f66779944da7364d3f5852572d01f42124fee998
a0473642bce536cc8960ffbb3dea88ff32cb75ec
/server.R
48be564253281b7d33d6af1134698ede0af7b8bf
[]
no_license
j11011/j11011.github.io
494ca020bc3163702f98f8cd05148b578714e2ec
4fd68de3e60facc2278328a3011dac7599a239a9
refs/heads/master
2020-03-12T03:50:39.726827
2018-04-28T02:34:24
2018-04-28T02:34:24
130,432,646
0
0
null
null
null
null
UTF-8
R
false
false
4,461
r
server.R
# # This is the server logic of a Shiny web application. You can run the # application by clicking 'Run App' above. # # Find out more about building applications with Shiny here: # # http://shiny.rstudio.com/ # library(shiny) library(shiny) library(shinyjs) library(scales) bsearch=function(nMin,nMax,nTarget){ nTarget=round(nTarget) text1=paste("Search from",nMin,"to",nMax,"the number",nTarget,":\n") text2="" arr=c(nMin:nMax) arr2=arr while(1){ if(nTarget>nMax | nTarget<nMin){ text1 = "Number not in array" break() } search1=round(mean(arr2)) if(search1==nTarget){ #print(paste("Encontre",nTarget)) text2=paste("Found",nTarget) text1=paste(text1,text2,"\n",sep="") break }else if(search1>nTarget){ nMax=search1-1 arr2=c(nMin:nMax) #print(paste(search1,"es mayor a", nTarget ,"ahora buscare de",nMin,"a",search1)) text2=paste(search1,"is greater than", nTarget ,"now I will look from",nMin,"to",nMax) }else if(search1<nTarget){ nMin=search1+1 arr2=c(nMin:nMax) #print(paste(search1,"es menor a", nTarget ,"ahora buscare de",search1,"a",nMax)) text2=paste(search1,"is less than", nTarget ,"now I will look from",nMin,"to",nMax) } text1=paste(text1,text2,"\n",sep="") } return(text1) } bubble_ord=function(arr){ text1="" #print(any(is.na(arr))) if(any(is.na(arr))==F){ for(i in (2:length(arr))){ ii=i-1 for(j in c(1:(length(arr)-ii))){ if(arr[j]>arr[j+1]){ aux=arr[j] arr[j]=arr[j+1] arr[j+1]=aux } text1=paste(text1,paste(arr,collapse=","),"\n") } } }else{ text1="The vector is wrong" } return(text1) } qs<-function(vec,start=1,finish=length(vec),text1="") { if(length(vec)>1){ if (finish>start) { pivot<-vec[start] N<-length(vec) window<-((1:N)>=start) & ((1:N)<=finish) low_part<-vec[(vec<pivot) & window] mid_part<-vec[(vec==pivot) & window] high_part<-vec[(vec>pivot) & window] if (start>1) text1=paste(text1,paste(vec[1:(start-1)],collapse = " "),"| ") text1=paste(text1,paste(low_part,collapse = " "),">>>", paste(mid_part,collapse = " "),"<<<",paste(high_part,collapse = " ")) #cat(low_part,">>>",mid_part,"<<<",high_part) if (finish<N) text1=paste(text1," |",paste(vec[(finish+1):N],collapse = " ")) text1=paste(text1,"\n") vec[window]<-c(low_part,mid_part,high_part) if (length(low_part)>0) { low_top<-start+length(low_part)-1 l_res<-qs(vec,start,low_top,text1=text1) text1=l_res[[2]] vec[start:low_top]=l_res[[1]][start:low_top] } if (length(high_part)>0) { high_bottom<-finish-length(high_part)+1 l_res<-qs(vec,high_bottom,finish,text1=text1) text1=l_res[[2]] vec[high_bottom:finish]=l_res[[1]][high_bottom:finish] } } }else{ vec=c(0) text1="" } return(list(vec,text1)) } matrix_frame=function(n){ if(n>1){ text=" 1" for(i in (2:n)){ text=paste0(text,",",toString(i)) } text=paste0(text,"\n") for(i in (1:n)){ text=paste0(text,toString(i)," ") for(i2 in (1:(n-1))){ text=paste0(text,",") #cat(text,"\n") } text=paste0(text,"\n") } }else{ text="Error!" } return(text) } borrarFirstn=function(n,text1){ for(i in (1:n)){ text1=gsub(paste0("\n",toString(i)),"\n",text1) } text1=gsub(" \n",",\n",text1) con <- textConnection(text1) data <- read.csv(con) #cat(text1) colnames(data)=c(1:ncol(data)) return(data) } # Define server logic required to draw a histogram shinyServer(function(input, output) { v <- reactiveValues(df_t= "0",df=data.frame()) #binary observeEvent(input$buscar1,{ nmin=isolate(as.numeric(input$min1)) nmax=isolate(as.numeric(input$max1)) ntarget=isolate(as.numeric(input$target1)) output$text1 <- renderText({bsearch(nmin,nmax,ntarget)}) }) #bubble observeEvent(input$ord2,{ arr2=as.numeric(unlist(strsplit(input$arr2, split=","))) output$text2 <- renderText({bubble_ord(arr2)}) }) #quicksort observeEvent(input$ord3,{ text1="" arr2=as.numeric(unlist(strsplit(input$arr3, split=","))) output$text3 <- renderText({qs(arr2,text1=text1)[[2]]}) }) })
54e5b17686b11c5cb33b6c80ebf9b8240fb79cd3
247946f5456e093a7fe49f57e722477ac9dc010e
/man/ezumap.Rd
d1cbce51e9456b2404e26cde5ad7d0ec68cd7b25
[ "MIT" ]
permissive
jdreyf/jdcbioinfo
b718d7e53f28dc15154d3a62b67075e84fbfa59b
1ce08be2c56688e8b3529227e166ee7f3f514613
refs/heads/master
2023-08-17T20:50:23.623546
2023-08-03T12:19:28
2023-08-03T12:19:28
208,874,588
3
1
null
null
null
null
UTF-8
R
false
true
3,099
rd
ezumap.Rd
% Generated by roxygen2: do not edit by hand % Please edit documentation in R/ezumap.R \name{ezumap} \alias{ezumap} \title{UMAP plot of first two dimensions} \usage{ ezumap( object, pheno.df, name = "umap", pca = TRUE, initial_dims = nrow(pheno.df), config = umap::umap.defaults, method = c("naive", "umap-learn"), preserve.seed = TRUE, alpha = 1, all.size = NULL, facet = NULL, title = NULL, subtitle = NULL, rm.leg.title = FALSE, labels = FALSE, manual.color = NULL, manual.shape = NULL, plot = TRUE, ... ) } \arguments{ \item{object}{Matrix-like object with features (e.g. genes) as rows and samples as columns.} \item{pheno.df}{Data frame with rows as samples and columns as phenotypes.} \item{name}{Name of file to create. Set to \code{NA} to plot to screen instead of to file.} \item{pca}{logical; Whether an initial PCA step should be performed (default: TRUE)} \item{initial_dims}{integer; the number of dimensions that should be retained in the initial PCA step (default: 50)} \item{config}{object of class umap.config} \item{method}{character, implementation. Available methods are 'naive' (an implementation written in pure R) and 'umap-learn' (requires python package 'umap-learn')} \item{preserve.seed}{logical, leave TRUE to insulate external code from randomness within the umap algorithms; set FALSE to allow randomness used in umap algorithms to alter the external random-number generator} \item{alpha}{Transparency, passed to \code{\link[ggplot2]{geom_point}}.} \item{all.size}{Passed to \code{\link[ggplot2]{geom_point}} \code{size} parameter to give size for all points without appearing in legend. \code{ggplot2} default is size=2.} \item{facet}{A formula with columns in \code{pheno.df} to facet by.} \item{title}{Title text; suppressed if it is \code{NULL}.} \item{subtitle}{Subtitle text; suppressed if it is \code{NULL} or \code{title} is \code{NULL}. If you'd like a \code{subtitle} but no \code{title}, set \code{title = ""}.} \item{rm.leg.title}{Logical indicating if legend title should be removed.} \item{labels}{Logical, should sample labels be added next to points?} \item{manual.color}{Vector passed to \code{\link[ggplot2:scale_manual]{scale_colour_manual}} for creating a discrete color scale. Vector length should be equal to number of levels in mapped variable.} \item{manual.shape}{Vector passed to \code{\link[ggplot2:scale_manual]{scale_shape_manual}} for creating a discrete color scale. Vector length should be equal to number of levels in mapped variable.} \item{plot}{Logical; should plot be generated?} \item{...}{list of settings with values overwrite defaults from UMAP \code{config} or passed to \code{\link[ggplot2:aes_]{aes_string}}.} } \value{ Invisibly, a \code{ggplot} object. Its \code{data} element contains the first two principal components appended to \code{pheno.df}. } \description{ UMAP plot of first two dimensions using \pkg{ggplot2}. } \details{ \code{object} must have colnames, and if \code{pheno.df} is given, it is checked that \code{colnames(object)==rownames(pheno.df)}. }
cdb4fd31f542405371b2519a8bbb5af44998a439
cc42b601ff02cf5b62007caef366249d0c430898
/misc/std12weeks.R
e8b5402a9885f6819568e62eb2b25063993b2603
[]
no_license
tkthomas27/r_code
a10bfa3f687b7dc002679571c8e23f2740345c6f
e6c15544e29942d47c94bfda497cf1ade47035fa
refs/heads/master
2020-04-05T11:44:52.349271
2018-07-06T19:35:42
2018-07-06T19:35:42
81,157,428
0
0
null
null
null
null
UTF-8
R
false
false
2,839
r
std12weeks.R
# *------------------------------------------------------------------ # | PROGRAM NAME: 12 week standard deviation # | DATE: 2017-12-07 # | CREATED BY: kyle thomas # *---------------------------------------------------------------- #set working directory setwd("/Volumes/tsandino_gomobile_project/Go Mobile/Paper_Daily Incentives_Select Stores/C. Stata files") # load libraries library(tidyverse) #data processing library(xts) # time series library(zoo) # time series library(padr) # adding missing days if desired # *---------------------------------------------------------------- # initialize data # *---------------------------------------------------------------- # read in data df <- read_csv("sales_variability_r.csv") # format date df$date <- as.Date(df$date, "%d %b %y") #split data by store dfs <- split(df,df$store) # *---------------------------------------------------------------- # calculate rolling 84 day weekly standard deviation # *---------------------------------------------------------------- # create empty dataframe to store the results store_results <- data.frame(date=NA, store=NA, sales = NA, std_12=NA) # loop through each store for(z in 1:35){ dfx <- dfs[[z]] #load select store #start with indices 1-84 and increment by 1 until none is left i<-1 j<-84 # create empty dataframe to store standard deviation results date = as.Date("2000-01-01") std_12 = 1111 std_results = data.frame(date,std_12) # if we want, we can pad out missing days #dfx <- pad(dfx) %>% replace_na(list(sales = 0)) # roll forward 84 day window by 1 and compute standard deviation on weekly basis while(j<nrow(dfx)){ #subset for 84 days df1 <- dfx[i:j,] #make into a time series object df2 <- xts(df1[,3], order.by = df1$date) #down sample to weekly by summing sales amounts weeks <- period.apply(df2, INDEX = endpoints(df2, on="days", k=7), FUN = sum) #compute standard deviation with anti-Bessel's correction std_12_week <- sqrt(sd(weeks$sales)^2 * (11/12)) #add to stored results std_results <- rbind(std_results, c(as.character(index(df2[84])), round(std_12_week,2))) #increment indices i<-i+1 j<-j+1 } #match store's std results to original data series; this may not be necessary df3 <- inner_join(dfx, std_results) #add to overall results file store_results <- rbind(store_results, df3) } # *---------------------------------------------------------------- # store results # *---------------------------------------------------------------- # drop blank first row store_results <- store_results[2:nrow(store_results),] # convert dates store_results$date <- as.Date(store_results$date) # write to csv write_csv(store_results,"store_results.csv")
9e680f7cd533fc3a971d4e609d25bf112cc7ae21
41503f4dbd79d3713c118ee79411b17b9025bfbf
/05-Famlia Apply.R
759243bb21b7d1b9fe9d7230ada3e8dfe705fe5a
[]
no_license
tcoliveira/R_Cap03
f795853e54ffd1bc2f0f6b30c58cb2062bb42aae
e3f8bebdda21181300624c70edd30501073eff67
refs/heads/master
2020-03-22T20:19:46.906283
2018-09-21T15:20:15
2018-09-21T15:20:15
140,590,377
0
0
null
null
null
null
UTF-8
R
false
false
1,616
r
05-Famlia Apply.R
##Vetores = cadeia ordenadas de elementos #Loops são ineficientes no R, pode ser a familia apply #Apply, aplica uma função a todas as linhas / colunas de uma matriz / df #lapply, retorna nova lista / sapply #tapply #Usando loop lista1 <- list(a = (1:10), b = (45:77)) ?sapply sapply(lista1, mean) ###objeto que quero percorrer, função. sapply é um loop x <- matrix(rnorm(9), nr = 3, byrow = T) x apply(x, 1, mean) apply(x, 2, mean) apply(x, 1, plot) resultapply <- apply(x, 1, mean) resultapply escola <- data.frame(Aluno = c("Allan", "Alice", "Aline", "Alana", "Alex", "Adovaldo"), Matematica = c(90,80,85,87,56,79), Geografia = c(88,81,85,20,21,30), Quimca = c(78,60,74,60,51,90)) escola escola$Geografia #Calculando a media por aluo escola$Media = NA escola escola$Media = apply(escola[,c(2,3,4)],1,mean) escola escola$Media = round(escola$Media) escola ##tapply() sqldf install.packages('sqldf') require(sqldf) escola2 <- data.frame(Aluno = c("Allan", "Alice", "Aline", "Alana", "Alex", "Adovaldo"), Semestre = c(1,1,1,2,2,2), Matematica = c(90,80,85,87,56,79), Geografia = c(88,81,85,20,21,30), Quimca = c(78,60,74,60,51,90)) escola2 sqldf("select aluno, sum(Matematica), sum(Geografia), sum(Quimca) from escola2 group by aluno") tapply(c(escola2$Matematica), escola2$Aluno, sum) ?by #lapply() ?lapply lista1<- list(a = (1:10), b = (45:77)) lista1 lapply(lista1, sum) sapply(lista1, sum) #vapply() vapply(lista1, fivenum, c(Min. =0, "1st"=0))
0ed1ba056f91ea94545cfe433af5cfa75051c393
c80d6dc4eb6a079fef9369095c1ac1bd0b8301d3
/R/lpUtil.R
92a1bbf2b517ff5acc410099663d6b8e458e0c6e
[]
no_license
pyrovski/MPItrace
5b674228277c4af666476fd584db2966e7cea16b
0645397426a5251becdd97c1c2474630501eb499
refs/heads/master
2020-05-17T03:16:34.837344
2015-03-19T01:16:25
2015-03-19T01:16:25
32,493,346
0
0
null
null
null
null
UTF-8
R
false
false
27,084
r
lpUtil.R
#!/usr/bin/env Rscript require('igraph') require('rjson') require('data.table') require('parallel') source('./util.R') source('./read.R') source('~/local/bin/pbutils.R') ilpFileTypes = c('duration', 'edges') fixedLPFileTypes = c('duration', 'edges') .solveLP = function(entrySpaceRow, nodePowerLimits, forceLocal=F, fixedOnly=T, roundMode='step'){ cuts = sub('.edges.csv','', Sys.glob(paste(gsub('[/. ]', '_', entrySpaceRow$key), '*.edges.csv', sep=''))) ilpCuts = grep('ILP', cuts, v=T) fixedCuts = setdiff(cuts, ilpCuts) rm(cuts) mclapply(fixedCuts, function(cut){ mclapply(nodePowerLimits, function(pl) { edgesFile = paste(cut, '.p', format(pl, nsmall=1), 'w.edges', sep='') cutEdges = paste(cut, '.edges.csv', sep='') if(file.exists(edgesFile) && file.info(edgesFile)$mtime > file.info(cutEdges)$mtime){ cat(edgesFile, 'exists\n') return(NULL) } command = paste('prefix=', cut, ' powerLimit=', pl*entrySpaceRow$ranks, if(forceLocal) ' FORCELOCAL=1', ' roundMode=', roundMode, ' ./fixed.sh', sep='') print(command) system(command, intern=T) }) }) } solveLP = function(...){ load('../mergedEntries.Rsave') mcrowApply(entrySpace, .solveLP, ...) } readLP = function(filename){ a = fromJSON(file=filename) f = function(b){ arraySel = grep('[[]', names(b)) singleSel = setdiff(1:length(b), arraySel) if(length(arraySel)){ arrayNames = gsub('[[].*[]]', '', names(b)[arraySel]) uArrayNames = unique(arrayNames) arrayVars = nnapply(uArrayNames, function(name){ b = b[grep(paste(name, '[[]', sep=''), names(b))] map = strsplit(gsub('[]]', '', names(b)), '[[]') indices = as.numeric(sapply(map, '[[', 2)) b = .rbindlist(lapply(b, as.data.table)) b$index = indices if('Id' %in% names(b)) b[, Id := NULL] b }) } else arrayVars = NULL if(length(singleSel)){ singleVars = lapply(b[singleSel], function(e){ b = as.data.table(e) if('Id' %in% names(b)) b[, Id := NULL] if(ncol(b) == 1 && identical(names(b), 'Value')) b = b[[1]] b }) } else singleVars = NULL c(arrayVars, singleVars) } a$Solution[[2]]$Variable = f(a$Solution[[2]]$Variable) a$Solution[[2]]$Constraint = f(a$Solution[[2]]$Constraint) a$Solution[[2]]$Objective = f(a$Solution[[2]]$Objective) a } #!@todo adapt for flow ilp, fixed ilp reconcileLP = function(resultFile, timesliceFile, powerLimit, mode='split'){ result = readLP(resultFile) vertexStartTimes = result$Solution[[2]]$Variable$vertexStartTime result$Solution[[2]]$Variable$vertexStartTime = NULL setnames(vertexStartTimes, c('index', 'Value'), c('vertex', 'start')) setkey(vertexStartTimes, vertex) tryCatch(load(timesliceFile), finally=NULL) vertices = slice[, list(vertex=union(src, dest))] setkey(vertices, vertex) vertices = vertexStartTimes[vertices] rm(vertexStartTimes) ##!@todo this is incorrect; some vertices are not present in this ##!timeslice or the next; just because they're destinations doesn't ##!mean we need to assign them start times. vertices[is.na(start), start := 0.0] e_uids = unique(slice[, e_uid]) taskDuration = result$Solution[[2]]$Variable$taskDuration setnames(taskDuration, c('index', 'Value'), c('e_uid', 'lpWeight')) setkey(taskDuration, e_uid) taskDuration = taskDuration[J(e_uids)] taskDuration[is.na(lpWeight), lpWeight := 0] taskPower = result$Solution[[2]]$Variable$taskPower setnames(taskPower, c('index', 'Value'), c('e_uid', 'lpPower')) setkey(taskPower, e_uid) taskPower = taskPower[J(e_uids)] setkey(taskPower, e_uid) ## these should not exist, at least for comp edges ##!@todo warn on NA power for comp edges taskPower = taskPower[slice[,head(.SD, 1),keyby=e_uid,.SDcols=c('type')]] if(nrow(taskPower[is.na(lpPower) & type == 'comp']) > 0){ stop('LP should provide all comp task power entries') } taskPower[is.na(lpPower) & type == 'comp', lpPower := 0] taskPower[, type := NULL] setkey(taskDuration, e_uid) task = merge(taskDuration, taskPower, all=T) rm(taskPower, taskDuration) result = list() if(mode=='keepAll'){ setkey(slice, e_uid) setkey(task, e_uid) ##edges = slice[task] result$edges = slice result$lp = task } else { ## mode != 'keepAll' setkey(slice, e_uid, weight, power) setkey(task, e_uid, lpWeight, lpPower) f = function(a, b) abs(a-b) < 1e-8 edges = lapply(e_uids, function(u){ s = slice[J(u)] if(nrow(s) == 1){ s[, frac := 1] return(s) } lp = task[J(u)] unconstrained = lp[lpWeight > .9 & (lpWeight %% 1) < .1] if(nrow(unconstrained) > 0){ cat('unconstrained weight(s)!\n') print(unconstrained) } ##!@todo this can be done with multiple e_uids at once ##! this needs to be approximate ##m = s[lp, nomatch=0] m = s[f(weight, lp$lpWeight) & f(power, lp$lpPower)] if(nrow(m) > 0){ m[, frac := 1] return(m) } ##!@todo figure out how to get Pyomo to be more precise with its output ##! can re-adjust lp weight based on selected power m = rbind(head(s[power < lp$lpPower], 1), tail(s[power > lp$lpPower], 1)) if(mode=='combined'){ ## find a single config that is closest to the LP m[, dist := sqrt(((power-lp$lpPower)/lp$lpPower)^2+((weight - lp$lpWeight)/lp$lpWeight)^2)] m = m[which.min(dist)] m$dist = NULL m$frac=1 return(m) } else if(mode == 'combinedLE'){ ## find a single config that is always under the power constraint m = m[power <= powerLimit, .SD[which.min(weight)], by=e_uid] m$frac=1 return(m) } else if(mode == 'split'){ ## split configs fastFrac = (lp$lpPower - m[1, power])/diff(m[, power]) slowFrac = 1 - fastFrac m$frac = c(fastFrac, slowFrac) ##! adjust weight by frac m[, weight := weight * frac] ###! m should contain two rows; one for each configuration neighboring ###! the LP-selected power/performance point return(m) } }) edges = .rbindlist(edges) result$edges = edges } result$vertices=vertices[order(start)] result } timeStr = '[0-9]+[.][0-9]+' ##!@todo save results from this function, check for newer inputs than previous result ##!@todo make sure result files are newer than csv and Rsave inputs readCommandResults = function(command, ...){ cat(command, '\n') resultFiles = list.files(pattern= paste(command, '_', timeStr, '[.]p.*w[.]results$', sep='')) powerLimits = unique(sub('w[.]results$', '', sub(paste(command, '_', timeStr, '[.]p', sep=''), '', resultFiles))) prefixes = unique(sub('[.]p.*w[.]results$', '', resultFiles)) timesliceFiles = paste(prefixes, '.Rsave', sep='') times = sub(paste(command, '_', sep=''), '', prefixes) f = function(powerLimit){ powerLimit = as.numeric(powerLimit) cat(powerLimit, 'w', '\n') resultFiles = list.files(pattern= paste(command, '_', timeStr, '[.]p', powerLimit, 'w[.]results$', sep='')) times = sub('[.]p.*w[.]results$', '', sub(paste(command, '_', sep=''), '', resultFiles)) result = mcmapply(reconcileLP, resultFiles, timesliceFiles, powerLimit, ..., SIMPLIFY=F) names(result) = times result } nnapply(powerLimits, f) } lpGo = function(...){ load('../mergedEntries.Rsave', envir=.GlobalEnv) files = list.files(pattern='.*[.]results$') commands <<- unique(sub(paste('_', timeStr, '[.]p.*w[.]results', sep=''),'',files)) nnapply(commands, readCommandResults, ...) } ilpGo = function(pattern='.*', powerLimitMin=0, ...){ ## get all files, then filter by prefix, then by power limit and cut load('../mergedEntries.Rsave', envir=.GlobalEnv) cutPattern = 'cut_[0-9]+' plPattern = 'p.*w' pattern = paste(pattern, '.*[.]duration$', sep='') files = list.files(pattern=pattern) ##duration' prefixes = unique(sub(paste('(.*)', cutPattern, plPattern, 'duration', sep='[.]'), '\\1', files)) if(!length(prefixes)){ warning('no ILP result files!', immediate.=T) return(NULL) } nnapply(prefixes, function(prefix){ fixed = length(grep('fixedLP', prefix)) > 0 files = list.files(pattern=paste(prefix, cutPattern, plPattern, 'duration$', sep='[.]')) powerLimits = sub('p([0-9.]+)w', '\\1', unique(sub(paste('.*', cutPattern, paste('(', plPattern, ')', sep=''), 'duration$', sep='[.]'), '\\1', files))) ##!@todo this needs to be modified to reformat the power limits with a trailing zero if(powerLimitMin > 0){ powerLimitFloat = as.numeric(powerLimits) cat('ignoring power limits: ', powerLimits[powerLimitFloat < powerLimitMin], '\n') powerLimits = powerLimits[powerLimitFloat >= powerLimitMin] } cuts = sort(as.numeric( sub('cut_([0-9]+)', '\\1', unique(sub(paste(prefix, paste('(',cutPattern, ')', sep=''), plPattern, 'duration$', sep='[.]'), '\\1', files))))) expectedCuts = as.integer(read.table(paste(prefix, '.cuts.csv', sep=''), h=F)[[1]]) nnapply(powerLimits, function(powerLimit){ plPattern = paste('p', powerLimit, 'w', sep='') presentCuts = list.files(pattern=paste(prefix, cutPattern, plPattern, 'edges$', sep='[.]')) presentCuts = as.integer(sub('.*cut_([0-9]+).*', '\\1', presentCuts)) if(length(setdiff(expectedCuts, presentCuts))){ errMsg = paste(prefix, '@', powerLimit, 'w:\nmissing cuts!\n', paste(setdiff(expectedCuts, presentCuts), collapse=' '), sep='') stop(errMsg) } nnapply(cuts, function(cut){ if(fixed) fileTypes = fixedLPFileTypes else fileTypes = ilpFileTypes nnapply(fileTypes, function(fileType){ filename = paste(prefix, paste('cut_', cut, sep=''), paste('p', powerLimit, 'w', sep=''), fileType, sep='.') tryCatch( as.data.table( read.table(filename, h=T, sep=',', strip.white=T)), error=function(e){ warning('failed to read ', filename, immediate.=T) NULL }, finally=NULL) } ) }, mc=T) } ) } ) } ##!@todo this function assumes that we don't alter the schedule from ##!the LP. For modes other than the default, this may not be true, ##!and we need to recompute start times and slack edges. lpMerge = function(slices, name){ edges = .rbindlist(napply(slices, function(e, name) { e$edges$ts = name e$edges }, mc=T)) vertices = .rbindlist(napply(slices, function(e, name) { e$vertices$ts =name e$vertices }, mc=T)) tsDuration = vertices[, .SD[which.max(start)], by=ts] tsDuration[, vertex := NULL] setnames(tsDuration, 'start', 'tsEnd') setkey(tsDuration, ts) tsDuration[, tsEnd := cumsum(tsEnd)] tsDuration$tsStart = 0 tsDuration$tsStart[2:nrow(tsDuration)] = head(tsDuration[, tsEnd], -1) setkey(vertices, ts) vertices = vertices[tsDuration[, list(ts, tsStart)]] vertices[, c('start', 'tsStart') := list(start + tsStart, NULL)] setnames(vertices, 'vertex', 'src') setkey(vertices, ts, src) setkey(edges, ts, src) edges = vertices[edges] setnames(vertices, 'src', 'vertex') ## renumber vertices across timeslices edges[, c('src', 'dest') := list(as.character(src), as.character(dest))] edges[, ts := as.character(.GRP), by=ts] edges[, ts := as.integer(ts)] edges[splitDest == T, dest := paste(src, '_', ts, 's', sep='')] edges[splitSrc == T, src := paste(src, '_', ts-1, 's', sep='')] edges[, c('splitSrc', 'splitDest') := list(NULL, NULL)] ## just to be consistent ##edges[splitSrc==F, src := paste(src, ts, sep='_')] ##edges[splitDest==F, dest := paste(dest, ts, sep='_')] ## assign new vertices to split-config edges from each timeslice, ## rename edge uids to be unique across timeslices if('frac' %in% names(edges)) edges = edges[order(ts, e_uid, -frac)] else edges = edges[order(ts, e_uid)] edges[, second := F] edges[, orig_e_uid := e_uid] edges = edges[,if(.N ==2){ e = copy(.SD) e[1, c('dest') := list(paste(src, '.', sep=''))] e[2, c('src', 'start', 'second') := list(paste(src, '.', sep=''), start + e[1, weight], T)] e } else { .SD }, by=list(e_uid, ts)] e_uid_map = data.table(orig=edges$e_uid) edges[, e_uid := as.character(e_uid)] edges[, e_uid := paste(ts, e_uid, sep='_')] edges[second == T, e_uid := paste(e_uid, '.', sep='')] e_uid_map$new = edges$e_uid ##!@todo assign weights to slack edges vertices = edges[, list(vertex=union(src, dest))] setkey(vertices, vertex) setkey(edges, src) vertices = vertices[unique(edges[,list(src, start)])] vertices = rbind(vertices, data.table(vertex='2', start=edges[dest=='2', max(start+weight)])) pt = powerTime(edges, vertices) plotPowerTime(pt, name=name) return(list(edges = edges, vertices = vertices, pt = pt)) } lpMergeAll = function(commands_powers){ napply(commands_powers, function(x, name){ command = name napply(x, function(x, name){ powerLimit = name lpMerge(x, name=paste(command, powerLimit)) }, mc=T) }) } loadAndMergeLP = function(){ results <<- lpGo() resultsMerged <<- lpMergeAll(results) resultsOneConf <<- lpGo(mode='combined') resultsMergedOneConf <<- lpMergeAll(resultsOneConf) resultsOneConfLE <<- lpGo(mode='combinedLE') resultsMergedOneConfLE <<- lpMergeAll(resultsOneConfLE) } # note: this also merges fixedLP results. I'm lazy. loadAndMergeILP = function(...){ resultsILP = ilpGo(...) ## retain only complete cuts ##!@todo get list of expected cuts, warn if any missing f = function(x, depth){ if(depth == 1){ if(any(sapply(x, is.null))) NULL else x } else { result = lapply(x, f, depth-1) result = result[!sapply(result, is.null)] if(length(result)) result else NULL } } resultsILP = f(resultsILP, 4) fixed = grep('fixedLP', names(resultsILP)) ilp = setdiff(seq(length(resultsILP)), fixed) resultsFixedLP = resultsILP[fixed] resultsILP = resultsILP[ilp] f = function(cuts, fileTypes) nnapply( fileTypes, function(fileType) .rbindlist( napply( cuts, function(x, name){ result=x[[fileType]] if(is.null(result)) return(result) result[, cut:=as.numeric(name)] result } ) ) ) ## merge cuts resultsILPMerged = lapply(resultsILP, lapply, f, ilpFileTypes) resultsFixedLPMerged = lapply(resultsFixedLP, lapply, f, fixedLPFileTypes) ## propagate event times across cuts. ##!@todo collectives are numbered in order of their occurrence, but ##!e.g. MPI_Waitall()s may not be. As the cut names are vertex ##!labels, we can get the vertex ordering from mergedData. ## f = function(x){ ## ### if we somehow avoid zero-length slack edges, the end time of a cut ## ### will not correspond to the start time of its last vertex ## ## place event start times within each cut ## setkey(x$duration, cut) ## x$duration[, cutEnd:=cumsum(duration)] ## x$duration[, cutStart:=c(0, head(cutEnd, -1))] ## setkey(x$events, cut, event) ## x$events = x$duration[x$events, list(event, cut, start=start+cutStart)] ## setkey(x$edges, cut, event) ## setkey(x$events, cut, event) ## x$edges = x$events[x$edges] ## ## renumber events, remove cut column ## x$activeEvents = x$edges[, list(event=unique(event)), by=cut] ## setkey(x$activeEvents, cut, event) ## x$activeEvents = x$activeEvents[, list(newEvent=.GRP-1), by=list(cut, event)] ## x$edges[x$activeEvents, c('event', 'cut') := list(newEvent, NULL)] ## x$events = x$events[x$activeEvents] ## x ## } ## # resultsILPMerged <<- lapply(resultsILPMerged, lapply, f) assign('resultsILPMerged', resultsILPMerged, envir=.GlobalEnv) assign('resultsFixedLPMerged', resultsFixedLPMerged, envir=.GlobalEnv) NULL } accumulateCutStarts = function(x, orderedCuts){ setkey(x$duration, cut) x$duration = x$duration[J(orderedCuts)] x$duration$cutEnd = as.numeric(NA) x$duration$cutStart = as.numeric(NA) x$duration[duration != 'infeasible', cutEnd:=cumsum(duration)] x$duration[duration != 'infeasible', cutStart:=c(0, head(cutEnd, -1))] x$duration[duration == 'infeasible', duration := as.numeric(NA)] x$duration$duration = as.numeric(x$duration$duration) setkey(x$duration, cut) setkey(x$edges, cut) x$edges = x$edges[x$duration[!is.na(duration), list(cut, cutStart)]] x$edges[, c('start', 'cutStart') := list(start+cutStart, NULL)] ## match NAs for message edges x$edges[power < 1, power := as.numeric(NA)] x } .writeILP_prefix = function(prefix){ origPrefix = sub('_fixedLP', '', prefix) eReduced = new.env() load(paste('../mergedData', origPrefix, 'Rsave', sep='.'), envir=eReduced) ranks = unique(eReduced$reduced$assignments$rank) ## eRuntimes = new.env() ## runtimes = ## load(paste('../', head(eReduced$reduced$assignments, 1)$date, ## '/merged.Rsave',sep=''), envir=eRuntimes) ##!@todo eReduced and eRuntimes have enough info to recreate the schedule? edges_inv = eReduced$reduced$edges_inv setkey(edges_inv, e_uid) vertices = eReduced$reduced$vertices rSched = eReduced$reduced$schedule globals = eReduced$reduced$globals rm(eReduced) ###!@todo this should be done in loadAndMergeILP, but requires ###!knowledge of vertex order that lives in merged_.Rsave. setkey(vertices, vertex) orderedCuts = vertices[J( resultsFixedLPMerged[[prefix]][[1]]$duration$cut )][order(start), vertex] resultsILPMerged = lapply(resultsILPMerged, lapply, accumulateCutStarts, orderedCuts) resultsFixedLPMerged = lapply(resultsFixedLPMerged, lapply, accumulateCutStarts, orderedCuts) assign('resultsILPMerged', resultsILPMerged, envir=.GlobalEnv) assign('resultsFixedLPMerged', resultsFixedLPMerged, envir=.GlobalEnv) cols = c('src', 's_uid', 'd_uid', 'dest', 'type', 'start', 'weight', ## name 'size', ## dest ## src 'tag', 'power', 'OMP_NUM_THREADS', 'cpuFreq') vertexCols = c('vertex', 'label', 'hash', 'reqs') .writeILP_prefix_powerLimit = function(pl){ if(!nrow(pl$edges)){ cat("powerLimit", pl$duration$powerLimit, ": no edges\n") return(NULL) } setkey(pl$edges, e_uid) ## merge sched with reduced edges_inv sched = edges_inv[pl$edges] save(sched, file= paste('sched', prefix, paste('p', pl$duration$powerLimit[1], 'w', sep=''), 'Rsave', sep='.')) .writePowerTime = function(s, label){ write.table(powerTime(s), file= paste('powerTime', prefix, label, paste('p', pl$duration$powerLimit[1], 'w', sep=''), 'dat', sep='.'), quote=F, sep='\t', row.names=F) } .writePowerTime(sched, 'all') lapply(ranks, function(r) .writePowerTime(sched[rank == r], label=sprintf('%06d', r))) ##!@todo plot per-rank power vs time, compare power allocation nonuniformity cols = intersect(cols, names(sched)) setkey(sched, src) schedDest = data.table::copy(sched) setkey(schedDest, dest) if(nrow(schedDest[J(2)]) != length(ranks)){ stop("MPI_Finalize anomaly found in LP schedule for prefix ", prefix, "!\n") } .writeILP_prefix_powerLimit_rank = function(r){ ## rank == dest rank for messages compEdges = vertices[, vertexCols, with=F][cbind( sched[type=='comp' & rank==r, cols, with=F], d_rank=as.integer(NA), s_rank=as.integer(NA), mseq=as.numeric(NA))] if(nrow(sched[type == 'message'])){ messageSendEdges = sched[type=='message' & s_rank==r] messageRecvEdges = sched[type=='message' & rank==r] messageSendEdges[, d_rank := rank] setkey(messageRecvEdges, 'o_dest') messageRecvEdges[, d_rank := rank] messageSendEdges = vertices[, vertexCols, with=F][messageSendEdges[, c(cols, 's_rank', 'd_rank'), with=F]] messageRecvEdges = vertices[, vertexCols, with=F][messageRecvEdges[, c(cols, 's_rank', 'd_rank', 'o_dest', 'o_d_uid'), with=F]] messageRecvEdges[, src := NULL] ### reduceConfs produces one message edge for each send/recv ### pair, but we want a separate row for both send and recv messageEdges = .rbindlist(list(messageRecvEdges, cbind(messageSendEdges, o_d_uid=as.numeric(NA)))) messageEdges[,mseq:=max(s_uid, o_d_uid, na.rm=T),by=vertex] messageEdges[, o_d_uid := NULL] edges = .rbindlist(list(compEdges, messageEdges)) rm(messageEdges) } else edges = compEdges edges[, mseq:=max(mseq, s_uid, na.rm=T), by=list(vertex,type)] edges = edges[, if(.N > 1){ ### we should never have more than one comp edge and one message edge ### leaving a vertex a = .SD[type=='comp'] a[, label := as.character(NA)] a = rbindlist(list(cbind(.SD[type=='message'], seq=1), cbind(a, seq=2))) ### hack to handle start times from sender in recv edges. a[, start := min(start)] } else { a=copy(.SD) a[,c('label', 'hash', 'reqs'):=as.character(NA)] cbind(rbindlist(list(.SD,a)), seq=1) }, by=vertex] edges[,mseq:=min(mseq),by=list(vertex)] edges = edges[order(mseq, seq)] ##!@todo UMT is missing MPI_Finalize; WTF? ## handle finalize edges = .rbindlist( list( edges, vertices[, vertexCols, with=F][cbind( schedDest[dest==2 & rank==r, cols, with=F], d_rank=as.integer(NA), s_rank=as.integer(NA), mseq=max(edges$mseq) + 1, seq=1)])) edges[, c('seq', 'vertex', 's_uid') := NULL] edges[, c('src', 'dest'):=as.integer(NA)] edges[type == 'message', c('src', 'dest') := list(as.integer(s_rank), as.integer(d_rank))] edges[, c('d_rank', 'type') := NULL] if(!'size' %in% names(edges)){ edges[, c('size', 'tag', 'comm') := as.numeric(NA)] } edges = edges[, list(start, duration=weight, name=sapply(strsplit(label, ' '), '[[', 1), size, dest, src, tag, comm='0x0', ##!@todo fix hash, flags=0, ##!@todo fix? #pkg_w=power, #pp0_w=0, #dram_w=0, reqs,##=as.character(NA), ##!@todo fix OMP_NUM_THREADS, cpuFreq )] for(col in setdiff(names(edges), c('reqs', 'name', 'comm', 'hash', 'pkg_w'))){ eCol = edges[[col]] eCol[is.na(eCol)] = globals$MPI_UNDEFINED edges[[col]] = eCol } edges[is.na(comm), comm:=0] edges[is.na(hash), hash:='0'] #edges[is.na(pkg_w), pkg_w:=0] edges[, cpuFreq:=as.integer(cpuFreq)] edges[!is.na(name), duration := 0.0] edges[, reqs:=sapply(reqs, paste, collapse=',')] write.table(edges, ## C code uses %s.%06d.dat file= paste('replay', prefix, paste('p', pl$duration$powerLimit[1], 'w', sep=''), sprintf('%06d', r), 'dat', sep='.'), quote=F, sep='\t', row.names=F) } ##debug(.writeILP_prefix_powerLimit_rank) lapply(ranks, .writeILP_prefix_powerLimit_rank) } mclapply(resultsFixedLPMerged[[prefix]], .writeILP_prefix_powerLimit) NULL } ## For each command, for each power limit, write a configuration ## schedule. This involves matching scheduled edges with edges from ## the original schedule, verifying that all edges were scheduled in ## the solution, matching edges with corresponding start vertices, ## writing vertices and edges in start order per rank, etc. We also ## require request IDs and communicator IDs. Perhaps it would be ## easier to load an existing replay schedule and add config options. writeILPSchedules = function(){ nnapply(names(resultsFixedLPMerged), .writeILP_prefix) } summarizeSchedules = function(){ napply( resultsFixedLPMerged, function(results, name){ prefix = name napply( results, function(plResults, name){ powerLimit = name powerTime = fread(paste('powerTime', prefix, 'all', paste('p', ###!@todo this should agree with .writePowerTime() as.integer(powerLimit), 'w', sep=''), 'dat', sep='.')) ## plot(stepfun(powerTime$start, c(powerTime$power,0))) meanPower = sum(powerTime[, diff(start) * tail(power, -1)])/tail(powerTime[, start],1) plResults$edges[, list(duration=max(start+weight), meanPower = meanPower, maxPower = max(powerTime$power))] })}) } if(!interactive()){ ## loadAndMergeLP() loadAndMergeILP() writeILPSchedules() }
90f2d0302c2534b6006593fdd4c918ac72708988
cec90ab7e7436b8ffabb7f5b5f38c7c7ee7904d2
/analysis/cSTM_time_dep_simulation.R
7f1227aae4b8dc185ee877710b49ab44cad05a96
[ "MIT" ]
permissive
DARTH-git/cohort-modeling-tutorial-timedep
d11df3e69f3f22e46bd4880031dd1564c0846c95
db0b7b20dbcccfd5a2676ce9e408bd19619be4b0
refs/heads/main
2023-07-21T04:28:22.968099
2023-07-13T19:00:42
2023-07-13T19:00:42
357,371,663
5
3
null
null
null
null
UTF-8
R
false
false
35,682
r
cSTM_time_dep_simulation.R
# Appendix code to Time-dependent cSTMs in R: Simulation-time dependency ---- #* This code forms the basis for the state-transition model of the tutorial: #* 'A Tutorial on Time-Dependent Cohort State-Transition Models in R using a #* Cost-Effectiveness Analysis Example' #* Authors: #* - Fernando Alarid-Escudero <falarid@stanford.edu> #* - Eline Krijkamp #* - Eva A. Enns #* - Alan Yang #* - M.G. Myriam Hunink #* - Petros Pechlivanoglou #* - Hawre Jalal #* Please cite the article when using this code #* #* To program this tutorial we used: #* R version 4.0.5 (2021-03-31) #* Platform: 64-bit operating system, x64-based processor #* Running under: Mac OS 12.2.1 #* RStudio: Version 1.4.1717 2009-2021 RStudio, Inc #******************************************************************************# # Description ---- #* This code implements a simulation-time-dependent Sick-Sicker cSTM model to #* conduct a CEA of four strategies: #* - Standard of Care (SoC): best available care for the patients with the #* disease. This scenario reflects the natural history of the disease #* progression. #* - Strategy A: treatment A is given to patients in the Sick and Sicker states, #* but only improves the quality of life of those in the Sick state. #* - Strategy B: treatment B is given to all sick patients and reduces disease #* progression from the Sick to Sicker state. #* - Strategy AB: This strategy combines treatment A and treatment B. The disease #* progression is reduced, and individuals in the Sick state have an improved #* quality of life. #******************************************************************************# # Initial setup ---- rm(list = ls()) # remove any variables in R's memory ## Install required packages ---- # install.packages("dplyr") # to manipulate data # install.packages("tidyr") # to manipulate data # install.packages("reshape2") # to manipulate data # install.packages("ggplot2") # to visualize data # install.packages("ggrepel") # to visualize data # install.packages("gridExtra") # to visualize data # install.packages("ellipse") # to visualize data # install.packages("scales") # for dollar signs and commas # install.packages(patchwork) # for combining ggplot2 figures # install.packages("dampack") # for CEA and calculate ICERs # install.packages("devtools") # to install packages from GitHub # devtools::install_github("DARTH-git/darthtools") # to install darthtools from GitHub using devtools # install.packages("doParallel") # to handle parallel processing ## Load packages ---- library(dplyr) library(tidyr) library(reshape2) # For melting data library(ggplot2) # For plotting library(ggrepel) # For plotting library(gridExtra) # For plotting library(ellipse) # For plotting library(scales) # For dollar signs and commas library(patchwork) # For combining ggplot2 figures # library(dampack) # Uncomment to use CEA and PSA visualization functionality from dampack instead of the functions included in this repository # library(darthtools) # Uncomment to use WCC, parameter transformation, and matrix checks from darthtools instead of the functions included in this repository # library(doParallel) # For running PSA in parallel ## Load supplementary functions ---- source("R/Functions.R") # Model input ---- ## General setup ---- cycle_length <- 1 # cycle length equal to one year (use 1/12 for monthly) n_age_init <- 25 # age at baseline n_age_max <- 100 # maximum age of follow up n_cycles <- (n_age_max - n_age_init)/cycle_length # time horizon, number of cycles #* Age labels v_age_names <- paste(rep(n_age_init:(n_age_max-1), each = 1/cycle_length), 1:(1/cycle_length), sep = ".") #* the 4 health states of the model: v_names_states <- c("H", # Healthy (H) "S1", # Sick (S1) "S2", # Sicker (S2) "D") # Dead (D) n_states <- length(v_names_states) # number of health states ### Discounting factors ---- d_c <- 0.03 # annual discount rate for costs d_e <- 0.03 # annual discount rate for QALYs ### Strategies ---- v_names_str <- c("Standard of care", # store the strategy names "Strategy A", "Strategy B", "Strategy AB") n_str <- length(v_names_str) # number of strategies ## Within-cycle correction (WCC) using Simpson's 1/3 rule ---- v_wcc <- gen_wcc(n_cycles = n_cycles, # Function included in "R/Functions.R". The latest version can be found in `darthtools` package method = "Simpson1/3") # vector of wcc ### Transition rates (annual), and hazard ratios (HRs) ---- r_HS1 <- 0.15 # constant annual rate of becoming Sick when Healthy r_S1H <- 0.5 # constant annual rate of becoming Healthy when Sick r_S1S2 <- 0.105 # constant annual rate of becoming Sicker when Sick hr_S1 <- 3 # hazard ratio of death in Sick vs Healthy hr_S2 <- 10 # hazard ratio of death in Sicker vs Healthy ### Effectiveness of treatment B ---- hr_S1S2_trtB <- 0.6 # hazard ratio of becoming Sicker when Sick under treatment B ## Age-dependent mortality rates ---- lt_usa_2015 <- read.csv("data/LifeTable_USA_Mx_2015.csv") #* Extract age-specific all-cause mortality for ages in model time horizon v_r_mort_by_age <- lt_usa_2015 %>% dplyr::filter(Age >= n_age_init & Age < n_age_max) %>% dplyr::select(Total) %>% as.matrix() ### State rewards ---- #### Costs ---- c_H <- 2000 # annual cost of being Healthy c_S1 <- 4000 # annual cost of being Sick c_S2 <- 15000 # annual cost of being Sicker c_D <- 0 # annual cost of being dead c_trtA <- 12000 # annual cost of receiving treatment A c_trtB <- 13000 # annual cost of receiving treatment B #### Utilities ---- u_H <- 1 # annual utility of being Healthy u_S1 <- 0.75 # annual utility of being Sick u_S2 <- 0.5 # annual utility of being Sicker u_D <- 0 # annual utility of being dead u_trtA <- 0.95 # annual utility when receiving treatment A ### Transition rewards ---- du_HS1 <- 0.01 # disutility when transitioning from Healthy to Sick ic_HS1 <- 1000 # increase in cost when transitioning from Healthy to Sick ic_D <- 2000 # increase in cost when dying ### Discount weight for costs and effects ---- v_dwc <- 1 / ((1 + (d_e * cycle_length)) ^ (0:n_cycles)) v_dwe <- 1 / ((1 + (d_c * cycle_length)) ^ (0:n_cycles)) # Process model inputs ---- ## Age-specific transition rates to the Dead state for all cycles ---- v_r_HDage <- rep(v_r_mort_by_age, each = 1/cycle_length) #* Name age-specific mortality vector names(v_r_HDage) <- v_age_names #* compute mortality rates v_r_S1Dage <- v_r_HDage * hr_S1 # Age-specific mortality rate in the Sick state v_r_S2Dage <- v_r_HDage * hr_S2 # Age-specific mortality rate in the Sicker state #* transform rates to probabilities adjusting by cycle length #* Function included in "R/Functions.R". The latest version can be found in `darthtools` package p_HS1 <- rate_to_prob(r = r_HS1, t = cycle_length) # constant annual probability of becoming Sick when Healthy conditional on surviving p_S1H <- rate_to_prob(r = r_S1H, t = cycle_length) # constant annual probability of becoming Healthy when Sick conditional on surviving p_S1S2 <- rate_to_prob(r = r_S1S2, t = cycle_length)# constant annual probability of becoming Sicker when Sick conditional on surviving v_p_HDage <- rate_to_prob(v_r_HDage, t = cycle_length) # Age-specific mortality risk in the Healthy state v_p_S1Dage <- rate_to_prob(v_r_S1Dage, t = cycle_length) # Age-specific mortality risk in the Sick state v_p_S2Dage <- rate_to_prob(v_r_S2Dage, t = cycle_length) # Age-specific mortality risk in the Sicker state ## Annual transition probability of becoming Sicker when Sick for treatment B ---- #* Apply hazard ratio to rate to obtain transition rate of becoming Sicker when #* Sick for treatment B r_S1S2_trtB <- r_S1S2 * hr_S1S2_trtB #* Transform rate to probability to become Sicker when Sick under treatment B #* adjusting by cycle length conditional on surviving #* (Function included in "R/Functions.R". The latest version can be found in #* `darthtools` package) p_S1S2_trtB <- rate_to_prob(r = r_S1S2_trtB, t = cycle_length) # Construct state-transition models ---- ## Initial state vector ---- #* All starting healthy v_m_init <- c(H = 1, S1 = 0, S2 = 0, D = 0) # initial state vector ## Initialize cohort traces ---- ### Initialize cohort trace under SoC ---- m_M_SoC <- matrix(NA, nrow = (n_cycles + 1), ncol = n_states, dimnames = list(0:n_cycles, v_names_states)) #* Store the initial state vector in the first row of the cohort trace m_M_SoC[1, ] <- v_m_init ### Initialize cohort trace for strategies A, B, and AB ---- #* Structure and initial states are the same as for SoC m_M_strA <- m_M_SoC # Strategy A m_M_strB <- m_M_SoC # Strategy B m_M_strAB <- m_M_SoC # Strategy AB ## Create transition probability arrays for strategy SoC ---- ### Initialize transition probability array for strategy SoC ---- #* All transitions to a non-death state are assumed to be conditional on survival a_P_SoC <- array(0, dim = c(n_states, n_states, n_cycles), dimnames = list(v_names_states, v_names_states, 0:(n_cycles - 1))) ### Fill in array ## From H a_P_SoC["H", "H", ] <- (1 - v_p_HDage) * (1 - p_HS1) a_P_SoC["H", "S1", ] <- (1 - v_p_HDage) * p_HS1 a_P_SoC["H", "D", ] <- v_p_HDage ## From S1 a_P_SoC["S1", "H", ] <- (1 - v_p_S1Dage) * p_S1H a_P_SoC["S1", "S1", ] <- (1 - v_p_S1Dage) * (1 - (p_S1H + p_S1S2)) a_P_SoC["S1", "S2", ] <- (1 - v_p_S1Dage) * p_S1S2 a_P_SoC["S1", "D", ] <- v_p_S1Dage ## From S2 a_P_SoC["S2", "S2", ] <- 1 - v_p_S2Dage a_P_SoC["S2", "D", ] <- v_p_S2Dage ## From D a_P_SoC["D", "D", ] <- 1 ### Initialize transition probability array for strategy A as a copy of SoC's ---- a_P_strA <- a_P_SoC ### Initialize transition probability array for strategy B ---- a_P_strB <- a_P_SoC #* Update only transition probabilities from S1 involving p_S1S2 a_P_strB["S1", "S1", ] <- (1 - v_p_S1Dage) * (1 - (p_S1H + p_S1S2_trtB)) a_P_strB["S1", "S2", ] <- (1 - v_p_S1Dage) * p_S1S2_trtB ### Initialize transition probability array for strategy AB as a copy of B's ---- a_P_strAB <- a_P_strB ## Check if transition probability arrays are valid ---- #* Functions included in "R/Functions.R". The latest version can be found in `darthtools` package ### Check that transition probabilities are [0, 1] ---- check_transition_probability(a_P_SoC, verbose = TRUE) check_transition_probability(a_P_strA, verbose = TRUE) check_transition_probability(a_P_strB, verbose = TRUE) check_transition_probability(a_P_strAB, verbose = TRUE) ### Check that all rows for each slice of the array sum to 1 ---- check_sum_of_transition_array(a_P_SoC, n_states = n_states, n_cycles = n_cycles, verbose = TRUE) check_sum_of_transition_array(a_P_strA, n_states = n_states, n_cycles = n_cycles, verbose = TRUE) check_sum_of_transition_array(a_P_strB, n_states = n_states, n_cycles = n_cycles, verbose = TRUE) check_sum_of_transition_array(a_P_strAB, n_states = n_states, n_cycles = n_cycles, verbose = TRUE) ## Create transition dynamics arrays ---- #* These arrays will capture transitions from each state to another over time ### Initialize transition dynamics array for strategy SoC ---- a_A_SoC <- array(0, dim = c(n_states, n_states, n_cycles + 1), dimnames = list(v_names_states, v_names_states, 0:n_cycles)) #* Set first slice of a_A_SoC with the initial state vector in its diagonal diag(a_A_SoC[, , 1]) <- v_m_init ### Initialize transition-dynamics array for strategies A, B, and AB ---- #* Structure and initial states are the same as for SoC a_A_strA <- a_A_SoC a_A_strB <- a_A_SoC a_A_strAB <- a_A_SoC # Run Markov model ---- #* Iterative solution of age-dependent cSTM for(t in 1:n_cycles){ ## Fill in cohort trace # For SoC m_M_SoC[t + 1, ] <- m_M_SoC[t, ] %*% a_P_SoC[, , t] # For strategy A m_M_strA[t + 1, ] <- m_M_strA[t, ] %*% a_P_strA[, , t] # For strategy B m_M_strB[t + 1, ] <- m_M_strB[t, ] %*% a_P_strB[, , t] # For strategy ZB m_M_strAB[t + 1, ] <- m_M_strAB[t, ] %*% a_P_strAB[, , t] ## Fill in transition-dynamics array # For SoC a_A_SoC[, , t + 1] <- diag(m_M_SoC[t, ]) %*% a_P_SoC[, , t] # For strategy A a_A_strA[, , t + 1] <- diag(m_M_strA[t, ]) %*% a_P_strA[, , t] # For strategy B a_A_strB[, , t + 1] <- diag(m_M_strB[t, ]) %*% a_P_strB[, , t] # For strategy AB a_A_strAB[, , t + 1] <- diag(m_M_strAB[t, ]) %*% a_P_strAB[, , t] } ## Store the cohort traces in a list ---- l_m_M <- list(SoC = m_M_SoC, A = m_M_strA, B = m_M_strB, AB = m_M_strAB) names(l_m_M) <- v_names_str ## Store the transition dynamics array for each strategy in a list ---- l_a_A <- list(SoC = a_A_SoC, A = a_A_strA, B = a_A_strB, AB = a_A_strAB) names(l_a_A) <- v_names_str # Plot Outputs ---- #* (Functions included in "R/Functions.R"; depends on the `ggplot2` package) ## Plot the cohort trace for strategy SoC ---- plot_trace(m_M_SoC) ## Plot the cohort trace for all strategies ---- plot_trace_strategy(l_m_M) ## Plot the epidemiology outcomes ---- ### Survival ---- survival_plot <- plot_surv(l_m_M, v_names_death_states = "D") + theme(legend.position = "bottom") survival_plot ### Prevalence ---- prevalence_S1_plot <- plot_prevalence(l_m_M, v_names_sick_states = c("S1"), v_names_dead_states = "D") + theme(legend.position = "") prevalence_S1_plot prevalence_S2_plot <- plot_prevalence(l_m_M, v_names_sick_states = c("S2"), v_names_dead_states = "D") + theme(legend.position = "") prevalence_S2_plot prevalence_S1S2_plot <- plot_prevalence(l_m_M, v_names_sick_states = c("S1", "S2"), v_names_dead_states = "D") + theme(legend.position = "") prevalence_S1S2_plot prop_sicker_plot <- plot_proportion_sicker(l_m_M, v_names_sick_states = c("S1", "S2"), v_names_sicker_states = c("S2")) + theme(legend.position = "bottom") prop_sicker_plot ## Combine plots ---- gridExtra::grid.arrange(survival_plot, prevalence_S1_plot, prevalence_S2_plot, prevalence_S1S2_plot, prop_sicker_plot, ncol = 1, heights = c(0.75, 0.75, 0.75, 0.75, 1)) # State Rewards ---- ## Scale by the cycle length ---- #* Vector of state utilities under strategy SoC v_u_SoC <- c(H = u_H, S1 = u_S1, S2 = u_S2, D = u_D) * cycle_length #* Vector of state costs under strategy SoC v_c_SoC <- c(H = c_H, S1 = c_S1, S2 = c_S2, D = c_D) * cycle_length #* Vector of state utilities under strategy A v_u_strA <- c(H = u_H, S1 = u_trtA, S2 = u_S2, D = u_D) * cycle_length #* Vector of state costs under strategy A v_c_strA <- c(H = c_H, S1 = c_S1 + c_trtA, S2 = c_S2 + c_trtA, D = c_D) * cycle_length #* Vector of state utilities under strategy B v_u_strB <- c(H = u_H, S1 = u_S1, S2 = u_S2, D = u_D) * cycle_length #* Vector of state costs under strategy B v_c_strB <- c(H = c_H, S1 = c_S1 + c_trtB, S2 = c_S2 + c_trtB, D = c_D) * cycle_length #* Vector of state utilities under strategy AB v_u_strAB <- c(H = u_H, S1 = u_trtA, S2 = u_S2, D = u_D) * cycle_length #* Vector of state costs under strategy AB v_c_strAB <- c(H = c_H, S1 = c_S1 + (c_trtA + c_trtB), S2 = c_S2 + (c_trtA + c_trtB), D = c_D) * cycle_length ## Store state rewards ---- #* Store the vectors of state utilities for each strategy in a list l_u <- list(SoC = v_u_SoC, A = v_u_strA, B = v_u_strB, AB = v_u_strAB) #* Store the vectors of state cost for each strategy in a list l_c <- list(SoC = v_c_SoC, A = v_c_strA, B = v_c_strB, AB = v_c_strAB) #* assign strategy names to matching items in the lists names(l_u) <- names(l_c) <- v_names_str # Compute expected outcomes ---- #* Create empty vectors to store total utilities and costs v_tot_qaly <- v_tot_cost <- vector(mode = "numeric", length = n_str) names(v_tot_qaly) <- names(v_tot_cost) <- v_names_str ## Loop through each strategy and calculate total utilities and costs ---- for (i in 1:n_str) { # i <- 1 v_u_str <- l_u[[i]] # select the vector of state utilities for the i-th strategy v_c_str <- l_c[[i]] # select the vector of state costs for the i-th strategy a_A_str <- l_a_A[[i]] # select the transition array for the i-th strategy ##* Array of state rewards #* Create transition matrices of state utilities and state costs for the i-th strategy m_u_str <- matrix(v_u_str, nrow = n_states, ncol = n_states, byrow = T) m_c_str <- matrix(v_c_str, nrow = n_states, ncol = n_states, byrow = T) #* Expand the transition matrix of state utilities across cycles to form a transition array of state utilities a_R_u_str <- array(m_u_str, dim = c(n_states, n_states, n_cycles + 1), dimnames = list(v_names_states, v_names_states, 0:n_cycles)) # Expand the transition matrix of state costs across cycles to form a transition array of state costs a_R_c_str <- array(m_c_str, dim = c(n_states, n_states, n_cycles + 1), dimnames = list(v_names_states, v_names_states, 0:n_cycles)) ##* Apply transition rewards #* Apply disutility due to transition from H to S1 a_R_u_str["H", "S1", ] <- a_R_u_str["H", "S1", ] - du_HS1 #* Add transition cost per cycle due to transition from H to S1 a_R_c_str["H", "S1", ] <- a_R_c_str["H", "S1", ] + ic_HS1 #* Add transition cost per cycle of dying from all non-dead states a_R_c_str[-n_states, "D", ] <- a_R_c_str[-n_states, "D", ] + ic_D ###* Expected QALYs and costs for all transitions per cycle #* QALYs = life years x QoL #* Note: all parameters are annual in our example. In case your own case example is different make sure you correctly apply. a_Y_c_str <- a_A_str * a_R_c_str a_Y_u_str <- a_A_str * a_R_u_str ###* Expected QALYs and costs per cycle ##* Vector of QALYs and costs v_qaly_str <- apply(a_Y_u_str, 3, sum) # sum the proportion of the cohort across transitions v_cost_str <- apply(a_Y_c_str, 3, sum) # sum the proportion of the cohort across transitions ##* Discounted total expected QALYs and Costs per strategy and apply within-cycle correction if applicable #* QALYs v_tot_qaly[i] <- t(v_qaly_str) %*% (v_dwe * v_wcc) #* Costs v_tot_cost[i] <- t(v_cost_str) %*% (v_dwc * v_wcc) } # Cost-effectiveness analysis (CEA) ---- ## Incremental cost-effectiveness ratios (ICERs) ---- #* Function included in "R/Functions.R"; depends on the `dplyr` package #* The latest version can be found in `dampack` package df_cea <- calculate_icers(cost = v_tot_cost, effect = v_tot_qaly, strategies = v_names_str) df_cea ## CEA table in proper format ---- table_cea <- format_table_cea(df_cea) # Function included in "R/Functions.R"; depends on the `scales` package table_cea ## CEA frontier ----- #* Function included in "R/Functions.R"; depends on the `ggplot2` and `ggrepel` packages. #* The latest version can be found in `dampack` package plot(df_cea, label = "all", txtsize = 16) + expand_limits(x = max(table_cea$QALYs) + 0.1) + theme(legend.position = c(0.8, 0.2)) #******************************************************************************# # Probabilistic Sensitivity Analysis (PSA) ----- ## Load model, CEA and PSA functions ---- source('R/Functions_cSTM_time_dep_simulation.R') source('R/Functions.R') ## List of input parameters ----- l_params_all <- list( # Transition probabilities (per cycle), hazard ratios v_r_HDage = v_r_HDage, # constant rate of dying when Healthy (all-cause mortality) r_HS1 = 0.15, # constant annual rate of becoming Sick when Healthy conditional on surviving r_S1H = 0.5, # constant annual rate of becoming Healthy when Sick conditional on surviving r_S1S2 = 0.105, # constant annual rate of becoming Sicker when Sick conditional on surviving hr_S1 = 3, # hazard ratio of death in Sick vs Healthy hr_S2 = 10, # hazard ratio of death in Sicker vs Healthy # Effectiveness of treatment B hr_S1S2_trtB = 0.6, # hazard ratio of becoming Sicker when Sick under treatment B ## State rewards # Costs c_H = 2000, # cost of remaining one cycle in Healthy c_S1 = 4000, # cost of remaining one cycle in Sick c_S2 = 15000, # cost of remaining one cycle in Sicker c_D = 0, # cost of being dead (per cycle) c_trtA = 12000, # cost of treatment A c_trtB = 13000, # cost of treatment B # Utilities u_H = 1, # utility when Healthy u_S1 = 0.75, # utility when Sick u_S2 = 0.5, # utility when Sicker u_D = 0, # utility when Dead u_trtA = 0.95, # utility when being treated with A ## Transition rewards du_HS1 = 0.01, # disutility when transitioning from Healthy to Sick ic_HS1 = 1000, # increase in cost when transitioning from Healthy to Sick ic_D = 2000, # increase in cost when dying # Initial and maximum ages n_age_init = 25, n_age_max = 100, # Discount rates d_c = 0.03, # annual discount rate for costs d_e = 0.03, # annual discount rate for QALYs, # Cycle length cycle_length = 1 ) #* Store the parameter names into a vector v_names_params <- names(l_params_all) ## Test functions to generate CE outcomes and PSA dataset ---- #* Test function to compute CE outcomes calculate_ce_out(l_params_all) # Function included in "R/Functions_cSTM_time_dep_simulation.R" #* Test function to generate PSA input dataset generate_psa_params(10) # Function included in "R/Functions_cSTM_time_dep_simulation.R" ## Generate PSA dataset ---- #* Number of simulations n_sim <- 1000 #* Generate PSA input dataset df_psa_input <- generate_psa_params(n_sim = n_sim) #* First six observations head(df_psa_input) ### Histogram of parameters ---- ggplot(melt(df_psa_input, variable.name = "Parameter"), aes(x = value)) + facet_wrap(~Parameter, scales = "free") + geom_histogram(aes(y = ..density..)) + ylab("") + theme_bw(base_size = 16) + theme(axis.text = element_text(size = 6), axis.title.x = element_blank(), axis.title.y = element_blank(), axis.text.y = element_blank(), axis.ticks.y = element_blank()) ## Run PSA ---- #* Initialize data.frames with PSA output #* data.frame of costs df_c <- as.data.frame(matrix(0, nrow = n_sim, ncol = n_str)) colnames(df_c) <- v_names_str #* data.frame of effectiveness df_e <- as.data.frame(matrix(0, nrow = n_sim, ncol = n_str)) colnames(df_e) <- v_names_str #* data.frame of survival m_surv <- matrix(NA, nrow = n_sim, ncol = (n_cycles + 1), dimnames = list(1:n_sim, 0:n_cycles)) df_surv <- data.frame(Outcome = "Survival", m_surv, check.names = "FALSE") #* data.frame of life expectancy df_le <- data.frame(Outcome = "Life Expectancy", LE = m_surv[, 1]) #* data.frame of prevalence of S1 m_prev <- matrix(NA, nrow = n_sim, ncol = (n_cycles + 1), dimnames = list(1:n_sim, 0:n_cycles)) df_prevS1 <- data.frame(States = "S1", m_prev, check.names = "FALSE") #* data.frame of prevalence of S2 df_prevS2 <- data.frame(States = "S2", m_prev, check.names = "FALSE") #* data.frame of prevalence of S1 & S2 df_prevS1S2 <- data.frame(States = "S1 + S2", m_prev, check.names = "FALSE") #* Conduct probabilistic sensitivity analysis #* Run Markov model on each parameter set of PSA input dataset n_time_init_psa_series <- Sys.time() for (i in 1:n_sim) { # i <- 1 l_psa_input <- update_param_list(l_params_all, df_psa_input[i,]) # Economics Measures l_out_ce_temp <- calculate_ce_out(l_psa_input) df_c[i, ] <- l_out_ce_temp$Cost df_e[i, ] <- l_out_ce_temp$Effect # Epidemiological Measures l_out_epi_temp <- generate_epi_measures_SoC(l_psa_input) df_surv[i, -1] <- l_out_epi_temp$S df_le[i, -1] <- l_out_epi_temp$LE df_prevS1[i, -1] <- l_out_epi_temp$PrevS1 df_prevS2[i, -1] <- l_out_epi_temp$PrevS2 df_prevS1S2[i, -1] <- l_out_epi_temp$PrevS1S2 # Display simulation progress if (i/(n_sim/100) == round(i/(n_sim/100), 0)) { # display progress every 5% cat('\r', paste(i/n_sim * 100, "% done", sep = " ")) } } n_time_end_psa_series <- Sys.time() n_time_total_psa_series <- n_time_end_psa_series - n_time_init_psa_series print(paste0("PSA with ", scales::comma(n_sim), " simulations run in series in ", round(n_time_total_psa_series, 2), " ", units(n_time_total_psa_series))) ## Run Markov model on each parameter set of PSA input dataset in parallel # Uncomment next section to run in parallel ## Get OS os <- get_os() no_cores <- parallel::detectCores() - 1 print(paste0("Parallelized PSA on ", os, " using ", no_cores, " cores.")) n_time_init_psa_parallel <- Sys.time() # ## Run parallelized PSA based on OS # if (os == "osx") { # # Initialize cluster object # cl <- parallel::makeForkCluster(no_cores) # # Register clusters # doParallel::registerDoParallel(cl) # # Run parallelized PSA # df_ce <- foreach::foreach(i = 1:n_sim, .combine = rbind) %dopar% { # l_out_temp <- calculate_ce_out(df_psa_input[i, ]) # df_ce <- c(l_out_temp$Cost, l_out_temp$Effect) # } # # Extract costs and effects from the PSA dataset # df_c <- df_ce[, 1:n_str] # df_e <- df_ce[, (n_str + 1):(2*n_str)] # # Register end time of parallelized PSA # n_time_end_psa_parallel <- Sys.time() # } # if (os == "windows") { # # Initialize cluster object # cl <- parallel::makeCluster(no_cores) # # Register clusters # doParallel::registerDoParallel(cl) # opts <- list(attachExportEnv = TRUE) # # Run parallelized PSA # df_ce <- foreach::foreach(i = 1:n_samp, .combine = rbind, # .export = ls(globalenv()), # .packages = c("dampack"), # .options.snow = opts) %dopar% { # l_out_temp <- calculate_ce_out(df_psa_input[i, ]) # df_ce <- c(l_out_temp$Cost, l_out_temp$Effect) # } # # Extract costs and effects from the PSA dataset # df_c <- df_ce[, 1:n_str] # df_e <- df_ce[, (n_str + 1):(2*n_str)] # # Register end time of parallelized PSA # n_time_end_psa_parallel <- Sys.time() # } # if (os == "linux") { # # Initialize cluster object # cl <- parallel::makeCluster(no_cores) # # Register clusters # doParallel::registerDoMC(cl) # # Run parallelized PSA # df_ce <- foreach::foreach(i = 1:n_sim, .combine = rbind) %dopar% { # l_out_temp <- calculate_ce_out(df_psa_input[i, ]) # df_ce <- c(l_out_temp$Cost, l_out_temp$Effect) # } # # Extract costs and effects from the PSA dataset # df_c <- df_ce[, 1:n_str] # df_e <- df_ce[, (n_str + 1):(2*n_str)] # # Register end time of parallelized PSA # n_time_end_psa_parallel <- Sys.time() # } # # Stop clusters # stopCluster(cl) # n_time_total_psa_parallel <- n_time_end_psa_parallel - n_time_init_psa_parallel # print(paste0("PSA with ", scales::comma(n_sim), " simulations run in series in ", # round(n_time_total_psa_parallel, 2), " ", # units(n_time_total_psa_parallel))) ## Visualize PSA results for CEA ---- ### Create PSA object ---- #* Function included in "R/Functions.R" The latest version can be found in `dampack` package l_psa <- make_psa_obj(cost = df_c, effectiveness = df_e, parameters = df_psa_input, strategies = v_names_str) l_psa$strategies <- v_names_str colnames(l_psa$effectiveness) <- v_names_str colnames(l_psa$cost) <- v_names_str #* Vector with willingness-to-pay (WTP) thresholds. v_wtp <- seq(0, 200000, by = 5000) ### Cost-Effectiveness Scatter plot ---- txtsize <- 13 #* Function included in "R/Functions.R"; depends on `tidyr` and `ellipse` packages. #* The latest version can be found in `dampack` package gg_scattter <- plot.psa(l_psa, txtsize = txtsize) + ggthemes::scale_color_colorblind() + ggthemes::scale_fill_colorblind() + scale_y_continuous("Cost (Thousand $)", breaks = number_ticks(10), labels = function(x) x/1000) + xlab("Effectiveness (QALYs)") + guides(col = guide_legend(nrow = 2)) + theme(legend.position = "bottom") gg_scattter ### Incremental cost-effectiveness ratios (ICERs) with probabilistic output ---- #* Compute expected costs and effects for each strategy from the PSA #* Function included in "R/Functions.R". The latest version can be found in `dampack` package df_out_ce_psa <- summary(l_psa) #* Function included in "R/Functions.R"; depends on the `dplyr` package #* The latest version can be found in `dampack` package df_cea_psa <- calculate_icers(cost = df_out_ce_psa$meanCost, effect = df_out_ce_psa$meanEffect, strategies = df_out_ce_psa$Strategy) df_cea_psa ### Plot cost-effectiveness frontier with probabilistic output ---- #* Function included in "R/Functions.R"; depends on the `ggplot2` and `ggrepel` packages. #* The latest version can be found in `dampack` package plot.icers(df_cea_psa, label = "all", txtsize = txtsize) + expand_limits(x = max(table_cea$QALYs) + 0.1) + theme(legend.position = c(0.8, 0.2)) ### Cost-effectiveness acceptability curves (CEACs) and frontier (CEAF) --- #* Functions included in "R/Functions.R". The latest versions can be found in `dampack` package ceac_obj <- ceac(wtp = v_wtp, psa = l_psa) #* Regions of highest probability of cost-effectiveness for each strategy summary(ceac_obj) #* CEAC & CEAF plot gg_ceac <- plot.ceac(ceac_obj, txtsize = txtsize, xlim = c(0, NA), n_x_ticks = 14) + ggthemes::scale_color_colorblind() + ggthemes::scale_fill_colorblind() + theme(legend.position = c(0.8, 0.48)) gg_ceac ### Expected Loss Curves (ELCs) ---- #* Function included in "R/Functions.R".The latest version can be found in `dampack` package elc_obj <- calc_exp_loss(wtp = v_wtp, psa = l_psa) elc_obj #* ELC plot gg_elc <- plot.exp_loss(elc_obj, log_y = FALSE, txtsize = txtsize, xlim = c(0, NA), n_x_ticks = 14, col = "full") + ggthemes::scale_color_colorblind() + ggthemes::scale_fill_colorblind() + # geom_point(aes(shape = as.name("Strategy"))) + scale_y_continuous("Expected Loss (Thousand $)", breaks = number_ticks(10), labels = function(x) x/1000) + theme(legend.position = c(0.4, 0.7),) gg_elc ### Expected value of perfect information (EVPI) ---- #* Function included in "R/Functions.R". The latest version can be found in `dampack` package evpi <- calc_evpi(wtp = v_wtp, psa = l_psa) #* EVPI plot gg_evpi <- plot.evpi(evpi, effect_units = "QALY", txtsize = txtsize, xlim = c(0, NA), n_x_ticks = 14) + scale_y_continuous("EVPI (Thousand $)", breaks = number_ticks(10), labels = function(x) x/1000) gg_evpi ### Combine all figures into one ---- patched_cea <- (gg_scattter + gg_ceac + plot_layout(guides = "keep"))/(gg_elc + gg_evpi) gg_psa_plots <- patched_cea + plot_annotation(tag_levels = 'A') gg_psa_plots ## Visualize PSA results for Epidemiological Measures ---- ### Wrangle PSA output ---- #* Combine prevalence measures df_prev <- dplyr::bind_rows(df_prevS1, df_prevS2, df_prevS1S2) #* Transform to long format df_surv_lng <- reshape2::melt(df_surv, id.vars = c("Outcome"), # value.name = "Survival", variable.name = "Time") df_prev_lng <- reshape2::melt(df_prev, id.vars = c("States"), # value.name = "Proportion", variable.name = "Time") #* Compute posterior-predicted 95% CI df_surv_summ <- data_summary(df_surv_lng, varname = "value", groupnames = c("Outcome", "Time")) df_le_summ <- data_summary(df_le, varname = "LE", groupnames = c("Outcome")) df_prev_summ <- data_summary(df_prev_lng, varname = "value", groupnames = c("States", "Time")) df_prev_summ$States <- ordered(df_prev_summ$States, levels = c("S1", "S2", "S1 + S2")) ### Plot epidemiological measures --- txtsize_epi <- 16 #### Survival --- gg_surv_psa <- ggplot(df_surv_summ, aes(x = as.numeric(Time), y = value, ymin = lb, ymax = ub)) + geom_line() + geom_ribbon(alpha = 0.4) + scale_x_continuous(breaks = number_ticks(8)) + xlab("Cycle") + ylab("Proportion alive") + theme_bw(base_size = txtsize_epi) + theme() gg_surv_psa #### Life Expectancy --- gg_le_psa <- ggplot(df_le, aes(x = LE)) + geom_density(color = "darkblue", fill = "lightblue") + scale_x_continuous(breaks = number_ticks(8)) + xlab("Life expectancy") + ylab("") + theme_bw(base_size = txtsize_epi) + theme( axis.text.y = element_blank(), axis.ticks = element_blank()) gg_le_psa #### Prevalence --- gg_prev_psa <- ggplot(df_prev_summ, aes(x = as.numeric(Time), y = value, ymin = lb, ymax = ub, color = States, linetype = States, fill = States)) + geom_line() + geom_ribbon(alpha = 0.4, color = NA) + scale_x_continuous(breaks = number_ticks(8)) + scale_y_continuous(breaks = number_ticks(8), labels = scales::percent_format(accuracy = 1)) + scale_color_discrete(name = "Health state", l = 50) + scale_linetype(name = "Health state") + scale_fill_discrete(name = "Health state", l = 50) + xlab("Cycle") + ylab("Prevalence (%)") + theme_bw(base_size = 16) + theme(legend.position = c(0.83, 0.83)) gg_prev_psa ### Combine all figures into one ---- patched_epi <- (gg_surv_psa / gg_le_psa) | gg_prev_psa gg_psa_epi_plots <- patched_epi + plot_annotation(tag_levels = 'A') gg_psa_epi_plots
e662f17cc6bbc65f6cabf7e63e7177e6514c22ac
a3011e4901d9cda7fa28b5c53079743273a34ab2
/extras/2020-04-13_draw-line-on-plot.R
7ef766c7e9bb5e30339364be3419d603cd8c6f13
[]
no_license
amit-agni/FiMATS
ee56427132989d2c0beacdf9be9225699a4c02b7
78d42982c7391f8fa9c80ab2ef2677202e8a4c44
refs/heads/master
2021-05-21T09:26:16.056174
2020-05-18T06:08:02
2020-05-18T06:08:02
252,637,321
0
0
null
null
null
null
UTF-8
R
false
false
1,063
r
2020-04-13_draw-line-on-plot.R
library(shiny,tidyverse) ui <- fluidPage( plotOutput("plot", width = "500px", height = "500px", hover=hoverOpts(id = "hover", delay = 1000, delayType = "throttle", clip = TRUE, nullOutside = TRUE), click="click")) server <- function(input, output, session) { plot <- ggplot(mtcars,aes(x=wt,y=qsec)) + geom_point() observe({ if(is.null(input$click$x)){ output$plot <- renderPlot({plot}) }else{ print(input$click$x) print(input$click$y) print(input$hover$x) print(input$hover$y) plot + geom_segment(aes(x=isolate(input$click$x) ,y=isolate(input$click$y) ,xend=isolate(input$hover$x) ,yend=isolate(input$hover$y))) } }) observeEvent(input$click,{ }) } shinyApp(ui, server)
35fe7b5e493ed23bae0b2d935fa1d2972df3bf5a
26f1cb213312ad204072dadd6b1163bcc0fa1bba
/exemples/chap7/7.09.R
58714c61dbdc7938a41bde9c14699bdb0b131638
[]
no_license
fmigone/livreR
a9b6f61a0aab902fb3b07fc49ea7dd642b65bdc3
998df678da1559ee03438c439335db796a416f2f
refs/heads/master
2020-04-22T05:47:14.083087
2018-05-01T14:39:10
2018-05-01T14:39:10
null
0
0
null
null
null
null
UTF-8
R
false
false
887
r
7.09.R
library(reshape2) # transposition # transposition : 1 ligne par logement, 1 colonne par date (pour que l'opération # ne soit pas trop lente on se limite à 1 million de lignes) # la mention d'une moyenne est factice car il n'y a qu'un # prix par date et par logement evol_prix <- dcast(calendar[1:1000000,], listing_id ~ date, fun = mean, value.var = "price") # calcul d'une moyenne par ligne (donc par logement) prix_par_logement <- data.frame(id=evol_prix$listing_id, moy=rowMeans(evol_prix[,-1], na.rm = TRUE)) # aperçu du résultat avec les 6 premières lignes head(prix_par_logement) # id moy # 1 9279 55.00000 # 2 11170 70.61056 # 3 11798 101.07735 # 4 11848 87.00000 # 5 14011 89.00000 # 6 17372 54.00000
0556cfaff07c48b4e6a07556e35770b37b427706
f66ea0275fc795a7d7caf39fb0f6e91fc5a66830
/man/yr_oldestplayer.Rd
b2f51ee3f875dc817c9f1daadfaa14b282fa8db1
[]
no_license
sunpark0405/nba
acc8910923f0a1b573956861e0c9fa49768ccf79
a882e4179e50f6bab620bd87e48e8b586777602f
refs/heads/main
2023-03-23T23:10:43.817658
2021-03-08T02:20:44
2021-03-08T02:20:44
341,405,412
0
0
null
null
null
null
UTF-8
R
false
true
537
rd
yr_oldestplayer.Rd
% Generated by roxygen2: do not edit by hand % Please edit documentation in R/yr_oldestplayer.R \name{yr_oldestplayer} \alias{yr_oldestplayer} \title{Year and Oldest Player Function} \usage{ yr_oldestplayer(year) } \arguments{ \item{year}{What year are we observing? Need to define.} } \description{ This function selects the oldest player in a given year. If there are multiple players with the same age, the function will return the top alphabetical name. } \examples{ yr_oldestplayer() } \keyword{age} \keyword{player} \keyword{year}
7e1b44d5d023cad8e0fef50e80e273ffdd1daa7e
12e0ddae06438b748d12a7f9c26e67cf682a8c16
/featureEngineering/setup_H2O.R
0b532785362a35b2ed3e72049d26be386c3996cc
[ "MIT" ]
permissive
christianadriano/ML_SelfHealingUtility
b05b2462c95a9aed9ac86af9e5eeb65bb07713d0
398ef99a7073c6383862fade85b8816e65a2fb1e
refs/heads/master
2021-10-07T20:45:51.281121
2018-12-05T09:16:05
2018-12-05T09:16:05
105,566,942
0
0
null
null
null
null
UTF-8
R
false
false
714
r
setup_H2O.R
### Setup H2O packages - http://h2o-release.s3.amazonaws.com/h2o/rel-wolpert/4/docs-website/h2o-r/docs/articles/getting_started.html #Remove any previously installed packages for R. if ("package:h2o" %in% search()) { detach("package:h2o", unload=TRUE) } if ("h2o" %in% rownames(installed.packages())) { remove.packages("h2o") } pkgs <- c("RCurl","jsonlite") for (pkg in pkgs) { if (! (pkg %in% rownames(installed.packages()))) { install.packages(pkg) } } #Download and install the latest H2O package for R. install.packages("h2o", type="source", repos=(c("http://h2o-release.s3.amazonaws.com/h2o/latest_stable_R"))) #Initialize H2O and run a demo to see H2O at work. library(h2o) h2o.init() demo(h2o.kmeans)
48941ad2ba3b3f12102510dae02777dbc0582f30
11318c146dc2d3e3a58bcd2874d6897d072ef3db
/plot3.R
206292b4c57e2ab9c68ce432f874bb9bb488275c
[]
no_license
Chesalov63/ExData_Plotting1
23faa1e8c651f86c100a507a9aad05c6ae2869f7
351d1b09907b19cedda4376c1043b3647b858177
refs/heads/master
2021-01-01T08:10:57.969343
2020-02-09T12:22:43
2020-02-09T12:22:43
239,191,440
0
0
null
2020-02-08T19:21:05
2020-02-08T19:21:04
null
UTF-8
R
false
false
1,617
r
plot3.R
## Start with downloading and unzipping UCI file filename <- "exdata_data_household_power_consumption.zip" if(!file.exists(filename)) { download.file( "https://archive.ics.uci.edu/ml/machine-learning-databases/00235/household_power_consumption.zip", filename) } unzip(filename) ## Read the data from dates 2007-02-01 and 2007-02-02 (2880 records) library(data.table) library(dplyr) filename <- "household_power_consumption.txt" headerDT <- fread(filename, sep = ";", na.strings = "?", nrows=0) powerDT <- fread(filename, sep = ";", na.strings = "?", skip = "1/2/2007", nrows=2880, col.names = colnames(headerDT)) ## Remove the original file, it's too big to keep it unzipped file.remove(filename) ## Set locale in case it differs from North-American usage Sys.setlocale("LC_TIME", "C") ## Create new column, because we need Date and Time in POSIXct format powerDT <- mutate(powerDT, DateTime = as.POSIXct(strptime(paste(Date, Time), "%d/%m/%Y %H:%M:%S") ) ) ## Construct the plot and save it to a PNG file (480x480 by default) png("plot3.png") with(powerDT, plot(Sub_metering_1 ~ DateTime, type = "l", ylab = "Energy sub metering", xlab = "")) with(powerDT, lines(DateTime, Sub_metering_2, col = "red")) with(powerDT, lines(DateTime, Sub_metering_3, col = "blue")) legend("topright", pch = 32, col = c("black", "blue", "red"), legend = c("Sub_metering_1", "Sub_metering_2", "Sub_metering_3"), lwd=1) dev.off()
9ea58d28cc11e6c1cef71bc893e4aec34aba21c9
4882312ae33d4555e654321b3fee4cf18a05d563
/man/methylomeStats.Rd
e5d55d0fbc4ad425129f6f8e0d01b8da05aac4fe
[]
no_license
timoast/methylQC
d3487c884b12e2a79544440b32aa48b738d8453b
fdba2f4917e5d442675780b455d5066ee33c8126
refs/heads/master
2020-09-21T20:15:40.768725
2016-09-13T08:33:52
2016-09-13T08:33:52
66,904,949
3
0
null
null
null
null
UTF-8
R
false
true
364
rd
methylomeStats.Rd
% Generated by roxygen2: do not edit by hand % Please edit documentation in R/methylomeStats.R \name{methylomeStats} \alias{methylomeStats} \title{Methylome statistics} \usage{ methylomeStats(data) } \arguments{ \item{data}{A dataframe} } \value{ a dataframe } \description{ Generate summary statistics for methylome data } \examples{ methylomeStats(methylome) }
c9a6ff6456fef3812f3f3a468ef1423c9dae8487
a8f1a464fdd18222c094ddeccb8da85c204393ca
/RNAseq_discovery_analysis.R
b6d599adb3ef64c7a74b2c4745a91b7b139301b1
[]
no_license
nikolayaivanov/PNET_MetPrediction
3d1e3f881ac5e85954907874934d3d191a3ff52b
4b9291beb8ec8b41408476a8a4af360c7c93bb69
refs/heads/main
2023-06-27T19:26:54.072462
2021-08-02T13:34:12
2021-08-02T13:34:12
379,664,521
0
0
null
null
null
null
UTF-8
R
false
false
18,977
r
RNAseq_discovery_analysis.R
# ########################################################################################### ##### Discovery DE analysis [localized vs metastatic] ########################################################################################### ###### Read in the data and peform differential expression analysis load('/athena/masonlab/scratch/users/nai2008/PNET_FinnertyProject/_rdas/discovery_dataset_metadata.rda') source('/athena/masonlab/scratch/users/nai2008/ivanov_functions.R') library(DESeq2) library(tximeta) # import data se = tximeta(coldata=discovery_dataset_metadata, type = "salmon") # found matching transcriptome: # [ Ensembl - Homo sapiens - release 97 ] # summarize transcript-level quantifications to gene-level gse = summarizeToGene(se) # get TPM matrix tpm = assays(gse)$abundance #get count matrix counts=assays(gse)$counts # make DESeqDataSet object dds = DESeqDataSet(gse, design = ~ Distant_Mets) #perform pre-filtering to keep only rows that have at least 10 reads total keep <- rowSums(counts(dds)) >= 10 dds <- dds[keep,] # make a transformed count matrix, using variance stabilizing transformation (VST) vsd = vst(dds, blind=FALSE) # run SVA (see chapter 8.1 of https://www.bioconductor.org/packages/devel/workflows/vignettes/rnaseqGene/inst/doc/rnaseqGene.html) library(sva) dds <- estimateSizeFactors(dds) # using 'avgTxLength' from assays(dds), correcting for library size dat <- counts(dds, normalized=TRUE) idx = rowMeans(dat) > 1 dat = dat[idx, ] mod = model.matrix(~ Distant_Mets, colData(dds)) mod0 <- model.matrix(~1, colData(dds)) svaobj = svaseq(dat, mod, mod0) # Number of significant surrogate variables is: 12 colnames(svaobj$sv)=paste0('SV_',1:ncol(svaobj$sv)) colData(dds) = as(cbind(as.data.frame(colData(gse)),svaobj$sv),'DataFrame') design(dds) = ~ SV_1 + SV_2 + SV_3 + SV_4 + SV_5 + SV_6 + SV_7 + SV_8 + SV_9 + SV_10 + SV_11 + SV_12 + Distant_Mets # examine how well the SVA method did at recovering the batch variable (i.e. which dataset the samples originated from) pdf('/athena/masonlab/scratch/users/nai2008/PNET_FinnertyProject/_pdfs/discovery_dataset_SVA_plots.pdf') par(mar=c(5.1,5.3,4.1,2.1)) x=factor(dds$Dataset) for(i in 1:12){ boxplot(svaobj$sv[, i] ~ x, xlab='Batch (Dataset)', ylab=paste0("SV", i), main=paste0("Surrogate Variable ", i), cex.main=2, cex.lab=2, cex.axis=1.5, outline=FALSE, col='lightgrey', ylim=c( min(as.vector(svaobj$sv[, i])), max(as.vector(svaobj$sv[, i])) ) ) points(as.vector(svaobj$sv[, i]) ~ jitter(as.numeric(x), amount=0.2), pch =21, col='black', bg='darkgrey', cex=1.4) } # for (i in 1:12) { # stripchart(svaobj$sv[, i] ~ dds$Dataset, vertical = TRUE, main = paste0("Surrogate Variable ", i), ylab=paste0("SV", i),xlab='Batch (Dataset)',cex.main=2, cex.lab=2, cex.axis=1.5) # abline(h = 0, lty='dashed') # } dev.off() # run DE analysis dds=DESeq(dds) which(is.na(mcols(dds)$betaConv)) # none # 93 rows did not converge in beta # omit rows that did not converge in beta (these are typically genes with very small counts and little power) # see https://support.bioconductor.org/p/65091/ ddsClean <- dds[which(mcols(dds)$betaConv),] # extract results rr=results(ddsClean, alpha=0.1, contrast=c('Distant_Mets','Y','N')) # contrast = c( the name of a factor in the design formula, name of the numerator level for the fold change, name of the denominator level for the fold change) summary(rr) # out of 29726 with nonzero total read count # adjusted p-value < 0.1 # LFC > 0 (up) : 184, 0.62% # LFC < 0 (down) : 219, 0.74% # outliers [1] : 0, 0% # low counts [2] : 0, 0% # (mean count < 0) # add gene symbols, chr, & Ensembl gene IDs library(AnnotationHub) hub = AnnotationHub() dm = query(hub, c("EnsDb", "sapiens", "97")) edb = dm[["AH73881"]] genes=as.data.frame(genes(edb)) mm=match(rownames(rr), genes$gene_id) length(which(is.na(mm))) # 0 rr$chr=as.vector(genes$seqnames[mm]) rr$Ensembl=as.vector(rownames(rr)) rr$gene=as.vector(genes$gene_name[mm]) # save releveant data save(se, gse, tpm, counts, ddsClean, rr, vsd, file='/athena/masonlab/scratch/users/nai2008/PNET_FinnertyProject/_rdas/discovery_dataset_DESeq2_DEbyMetStatus_BUNDLE.rda') # load('/athena/masonlab/scratch/users/nai2008/PNET_FinnertyProject/_rdas/discovery_dataset_DESeq2_DEbyMetStatus_BUNDLE.rda') # se, gse, tpm, counts, ddsClean, rr, vsd ########################################################################################### ##### Downstream analysis ########################################################################################### library(DESeq2) source('/athena/masonlab/scratch/users/nai2008/ivanov_functions.R') load('/athena/masonlab/scratch/users/nai2008/PNET_FinnertyProject/_rdas/discovery_dataset_metadata.rda') load('/athena/masonlab/scratch/users/nai2008/PNET_FinnertyProject/_rdas/discovery_dataset_DESeq2_DEbyMetStatus_BUNDLE.rda') # se, gse, tpm, counts, ddsClean, rr, vsd ################ PCA # pcaData <- plotPCA(vsd, intgroup = c( "Distant_Mets"), returnData = TRUE) # percentVar <- round(100 * attr(pcaData, "percentVar")) library(calibrate) pca = prcomp(t(assays(vsd)[[1]])) #command that will return % of variance explained by each PC: pcaVars=getPcaVars(pca) all(rownames(pca$x) == discovery_dataset_metadata$names) #TRUE PCs=as.matrix(pca$x) ## PCA colored by metastatic status col=discovery_dataset_metadata$Distant_Mets col=sub('N','forestgreen',col) col=sub('Y','red',col) pdf('/athena/masonlab/scratch/users/nai2008/PNET_FinnertyProject/_pdfs/discovery_datset_PCA_plots_metStatus.pdf') par(mar=c(5.1,5.3,4.1,2.1)) for(i in seq(from=1, to=10, by=2)){ x_lab=paste0('PC',i,': ',signif(pcaVars[i],2),'% variance') y_lab=paste0('PC',i+1,': ',signif(pcaVars[i+1],2),'% variance') plot(PCs[,i], PCs[,i+1], xlab=x_lab, ylab=y_lab, pch=21, cex=1.2, col='black', bg=col, cex.lab=2, cex.axis=2, xlim=c(min(PCs[,i])-10,max(PCs[,i])+10)) textxy(PCs[,i], PCs[,i+1],discovery_dataset_metadata$names,cex =.7, offset = .7) legend("topright", legend=c('Distant mets','No distant mets'), pch=15, col=c("red", "forestgreen"), cex=1, pt.cex=1) } dev.off() pdf('/athena/masonlab/scratch/users/nai2008/PNET_FinnertyProject/_pdfs/discovery_datset_PCA_plots_metStatus_only_interesting_samples_labeled.pdf') par(mar=c(5.1,5.3,4.1,2.1)) for(i in seq(from=1, to=10, by=2)){ x_lab=paste0('PC',i,': ',signif(pcaVars[i],2),'% variance') y_lab=paste0('PC',i+1,': ',signif(pcaVars[i+1],2),'% variance') plot(PCs[,i], PCs[,i+1], xlab=x_lab, ylab=y_lab, pch=21, cex=1.2, col='black', bg=col, cex.lab=2, cex.axis=2, xlim=c(min(PCs[,i])-10,max(PCs[,i])+10)) textxy(PCs[c(15,22),i], PCs[c(15,22),i+1],discovery_dataset_metadata$names[c(15,22)],cex =.7, offset = .7) legend("topright", legend=c('Distant mets','No distant mets'), pch=15, col=c("red", "forestgreen"), cex=1, pt.cex=1) } dev.off() ## PCA colored by study col=c('blue','darkred') palette(col) study=as.factor(discovery_dataset_metadata$Dataset) pdf('/athena/masonlab/scratch/users/nai2008/PNET_FinnertyProject/_pdfs/discovery_datset_PCA_plots_studyID.pdf') par(mar=c(5.1,5.3,4.1,2.1)) for(i in seq(from=1, to=10, by=2)){ x_lab=paste0('PC',i,': ',signif(pcaVars[i],2),'% variance') y_lab=paste0('PC',i+1,': ',signif(pcaVars[i+1],2),'% variance') plot(PCs[,i], PCs[,i+1], xlab=x_lab, ylab=y_lab, pch=21, cex=1.2, col='black', bg=study, cex.lab=2, cex.axis=2, xlim=c(min(PCs[,i])-10,max(PCs[,i])+10)) #textxy(PCs[,i], PCs[,i+1],discovery_dataset_metadata$names,cex =.7, offset = .7) legend("topright", legend=levels(study), pch=15, col=col, cex=1, pt.cex=1, title='Study') } dev.off() ################ Make table of DE genes # print results sig_results=as.data.frame(rr[which(rr$padj<=0.1),]) # order results by LFC oo=order(sig_results$log2FoldChange) sig_results=sig_results[oo,] # make output table out=data.frame(gene=sig_results$gene, chr=sig_results$chr, Ensembl=sig_results$Ensembl, log2FoldChange=sig_results$log2FoldChange, FDR=sig_results$padj) nrow(out) # 403 length(which(out$log2FoldChange > 0)) # 184 genes overexpressed in samples with distant mets (relative to localized samples) length(which(out$log2FoldChange < 0)) # 219 genes are underexpressed in samples with distant mets (relative to localized samples) # How many DE genes are TFs? out$TF=FALSE TFs=read.csv("/athena/masonlab/scratch/users/nai2008/Human_TFs_DatabaseExtract_v_1.01.csv") TFs$Ensembl_ID=as.vector(TFs$Ensembl_ID) nrow(TFs) # 2765 length(unique(TFs$Ensembl_ID)) # 2765 which(is.na(TFs$Ensembl_ID)) # 0 mm=match(out$Ensembl,TFs$Ensembl_ID) which(duplicated(out$Ensembl)) # none which(duplicated(TFs$Ensembl_ID)) # none out$TF[which(!is.na(mm))]=TRUE length(which(out$TF==TRUE)) # 26 save(out, file='/athena/masonlab/scratch/users/nai2008/PNET_FinnertyProject/_rdas/discovery_dataset_DE_genes_byMetStatus.rda') # load('/athena/masonlab/scratch/users/nai2008/PNET_FinnertyProject/_rdas/discovery_dataset_DE_genes_byMetStatus.rda') # print table of DE genes write.csv(out,file="/athena/masonlab/scratch/users/nai2008/PNET_FinnertyProject/_data_tables/discovery_dataset_DE_genes_byMetStatus.csv", row.names=FALSE) ################ Perform gene set over-representation analysis (ORA) library(goseq) load('/athena/masonlab/scratch/users/nai2008/items_for_goseq_analysis.rda') # gene2cat_GOandKEGG, KEGG_term_names, median_tx_lengths, cat2gene_GO, cat2gene_KEGG if(length(which(is.na(rr$padj))) != 0) { rr_mod=rr[-which(is.na(rr$padj)),] } else { rr_mod = rr } indicator=rep(0, times=nrow(rr_mod)) indicator[which(rr_mod$padj<=0.1)]=1 aa=indicator names(aa)=rr_mod$Ensembl mm = match(names(aa), median_tx_lengths$gene_EnsemblID) bias.data = median_tx_lengths$median_length[mm] pwf = nullp(aa, 'hg38', 'ensGene', bias.data = bias.data) pdf('/athena/masonlab/scratch/users/nai2008/PNET_FinnertyProject/_pdfs/discovery_dataset_goseq_pwf_plot.pdf') plotPWF(pwf) dev.off() GO.KEGG.wall=goseq(pwf,"hg38","ensGene", gene2cat = gene2cat_GOandKEGG, test.cats=c("GO:CC", "GO:BP", "GO:MF", "KEGG")) GO.KEGG.wall$over_represented_FDR=p.adjust(GO.KEGG.wall$over_represented_pvalue, method="BH") GO.KEGG.wall$ontology[grep('path:hsa', GO.KEGG.wall$category)]='KEGG' index = grep('path:hsa', GO.KEGG.wall$category) for (i in 1:length(index)){ mm=match(GO.KEGG.wall$category[index[i]], KEGG_term_names$KEGG_ID) GO.KEGG.wall$term[index[i]] = KEGG_term_names$KEGG_term[mm] } length(which(GO.KEGG.wall$over_represented_FDR<=0.1)) # 2 GO.KEGG.wall_sig = GO.KEGG.wall[which(GO.KEGG.wall$over_represented_FDR<=0.1),] # Add DE genes in each GO/KEGG category GO.KEGG.wall_sig_withoutGenes = GO.KEGG.wall_sig library(AnnotationHub) hub = AnnotationHub() dm = query(hub, c("EnsDb", "sapiens", "97")) edb = dm[["AH73881"]] genes=as.data.frame(genes(edb)) ens.gene.map = data.frame(gene_id=genes$gene_id, gene_name=genes$gene_name) length(names(cat2gene_GO)) == length(unique(names(cat2gene_GO))) #TRUE length(names(cat2gene_KEGG)) == length(unique(names(cat2gene_KEGG))) #TRUE GO.KEGG.wall_sig$genes_Ensembl=NA GO.KEGG.wall_sig$genes=NA for (i in 1:nrow(GO.KEGG.wall_sig)){ cat=GO.KEGG.wall_sig$category[i] if (length(grep('GO',cat)) == 1){ m.cat=match(cat, names(cat2gene_GO)) if(is.na(m.cat)){print('error: m.cat does not match (GO)')} else { possible_genes=cat2gene_GO[[m.cat]] m.genes=match(possible_genes,names(aa)) if( length(which(!is.na(m.genes)))==0 ){print('error: m.genes are all <NA> (GO)')} else { if (length(which(is.na(m.genes)))>0){ possible_genes= possible_genes[-which(is.na(m.genes))] } m.genes=match(possible_genes,names(aa)) subset=aa[m.genes] DE_genes=subset[which(subset==1)] GO.KEGG.wall_sig$genes_Ensembl[i]=paste(names(DE_genes),collapse=';') m.ens=match(names(DE_genes),ens.gene.map$gene_id) GO.KEGG.wall_sig$genes[i]=paste(ens.gene.map$gene_name[m.ens],collapse=';') } } } else if (length(grep('path:hsa',cat)) == 1){ m.cat=match(cat, names(cat2gene_KEGG)) if(is.na(m.cat)){print('error: m.cat does not match (KEGG)')} else { possible_genes=cat2gene_KEGG[[m.cat]] m.genes=match(possible_genes,names(aa)) if(length(which(!is.na(m.genes) == 0))){print('error: m.genes are all <NA> (KEGG)')} else{ if (length(which(is.na(m.genes)))>0){ possible_genes= possible_genes[-which(is.na(m.genes))] } m.genes=match(possible_genes,names(aa)) subset=aa[m.genes] DE_genes=subset[which(subset==1)] GO.KEGG.wall_sig$genes_Ensembl[i]=paste(names(DE_genes),collapse=';') m.ens=match(names(DE_genes),ens.gene.map$gene_id) GO.KEGG.wall_sig$genes[i]=paste(ens.gene.map$gene_name[m.ens],collapse=';') } } } } write.csv(GO.KEGG.wall_sig, file = '/athena/masonlab/scratch/users/nai2008/PNET_FinnertyProject/_data_tables/discovery_dataset_DEbyMetStatus_ORA.csv') ################ Make TPM plots of validated genes with concordant LFCs b/w the discovery and validation datasets load('/athena/masonlab/scratch/users/nai2008/PNET_FinnertyProject/_rdas/validated_genes_DEbyMetStatus_concordantLFC_only.rda') log2_tpm_plus1=log2(tpm+1) x=as.vector(ddsClean$Distant_Mets) x=sub('Y','Distant mets',x) x=sub('N','No distant mets',x) x=factor(x, levels=c('No distant mets','Distant mets')) mycol=as.vector(ddsClean$Distant_Mets) mycol[which(mycol=="N")]='purple' mycol[which(mycol=="Y")]='green' pdf('/athena/masonlab/scratch/users/nai2008/PNET_FinnertyProject/_pdfs/discovery_dataset_TPM_plots_of_validated_genes_with_concordantLFCs.pdf') for (i in 1: nrow(validated_genes_concordantLFC_only)){ par(mar=c(5.1,5.3,4.1,2.1)) index=which(rr$Ensembl==validated_genes_concordantLFC_only$Ensembl[i]) zag1=paste0(rr$gene[index], ' (',rr$Ensembl[index],')') zag2=as.expression(bquote(log[2]~"FC" == .(signif(rr$log2FoldChange[index],2)))) zag3=paste0("FDR = ",signif(rr$padj[index],2)) log2_tpm_plus1_subset=as.vector(log2_tpm_plus1[which(rownames(log2_tpm_plus1)==validated_genes_concordantLFC_only$Ensembl[i]),]) boxplot(as.vector(log2_tpm_plus1_subset)~x, , xlab='Tumor status', ylab= as.expression(bquote(log[2]~"(TPM+1)")), main=zag1, cex.main=2, cex.lab=2, cex.axis=1.5, outline=FALSE, col='lightgrey', ylim=c( min(as.vector(log2_tpm_plus1_subset)), max(as.vector(log2_tpm_plus1_subset)) ) ) points(as.vector(log2_tpm_plus1_subset) ~ jitter(as.numeric(x), amount=0.2), pch =21, col='black', bg=mycol, cex=1.4) legend(x='topright',legend=c(zag2,zag3), bty='n') } dev.off() ################ UCHL1 expression #UCHL1 is ENSG00000154277 rr[which(rr$Ensembl=='ENSG00000154277'),] # baseMean log2FoldChange lfcSE stat pvalue # <numeric> <numeric> <numeric> <numeric> <numeric> # ENSG00000154277 18982.2 -0.297401 0.559588 -0.531464 0.595097 # padj chr Ensembl gene # <numeric> <character> <character> <character> # ENSG00000154277 0.999998 4 ENSG00000154277 UCHL1 ################ Clustering analysis load('/athena/masonlab/scratch/users/nai2008/PNET_FinnertyProject/_rdas/validated_genes_DEbyMetStatus_concordantLFC_only.rda') load('/athena/masonlab/scratch/users/nai2008/PNET_FinnertyProject/_rdas/discovery_dataset_metadata.rda') # Patients from whom samples CA20 and CA35 were derived had localized disease at the time of sequencing, but later developed distant mets; add that info to the metadata discovery_dataset_metadata$Loc.to.Met=NA discovery_dataset_metadata$Loc.to.Met[which(discovery_dataset_metadata$Distant_Mets=='Y')]=NA discovery_dataset_metadata$Loc.to.Met[which(discovery_dataset_metadata$names=='CA20')]='Yes' discovery_dataset_metadata$Loc.to.Met[which(discovery_dataset_metadata$names=='CA35')]='Yes' ###### Read in the data source('/athena/masonlab/scratch/users/nai2008/ivanov_functions.R') library(DESeq2) library(tximeta) # import data se = tximeta(coldata=discovery_dataset_metadata, type = "salmon") # found matching transcriptome: # [ Ensembl - Homo sapiens - release 97 ] # summarize transcript-level quantifications to gene-level gse = summarizeToGene(se) # get TPM matrix tpm = assays(gse)$abundance #get count matrix counts=assays(gse)$counts # make DESeqDataSet object dds = DESeqDataSet(gse, design = ~ Distant_Mets) #perform pre-filtering to keep only rows that have at least 10 reads total keep <- rowSums(counts(dds)) >= 10 dds <- dds[keep,] # make a transformed count matrix, using variance stabilizing transformation (VST) vsd = vst(dds, blind=FALSE) vst_counts = as.matrix(assay(vsd)) # regress out the batch variable library(jaffelab) mod = model.matrix(~Distant_Mets + factor(Dataset), data=as.data.frame(colData(dds))) clean_vst_counts=cleaningY(vst_counts, mod, P=2) library(AnnotationHub) hub = AnnotationHub() dm = query(hub, c("EnsDb", "sapiens", "97")) edb = dm[["AH73881"]] genes=as.data.frame(genes(edb)) mm_validated_genes_w_concordantLFCs=match(validated_genes_concordantLFC_only$Ensembl, rownames(clean_vst_counts)) if(length(which(is.na(mm_validated_genes_w_concordantLFCs))) != 0) { validated_genes_concordantLFC_only = validated_genes_concordantLFC_only[-which(is.na(mm_validated_genes_w_concordantLFCs)),] mm_validated_genes_w_concordantLFCs=match(validated_genes_concordantLFC_only$Ensembl, rownames(clean_vst_counts)) } which(duplicated(validated_genes_concordantLFC_only$Ensembl))# none which(duplicated(rownames(clean_vst_counts)))# none mm=match(rownames(clean_vst_counts),genes$gene_id) which(duplicated(rownames(clean_vst_counts)))# none which(duplicated(genes$gene_id))# none which(is.na(mm)) # none rownames(clean_vst_counts) = genes$gene_name[mm] all(colnames(clean_vst_counts)==discovery_dataset_metadata$names) #TRUE library(pheatmap) library(RColorBrewer) annotation_col=data.frame(Distant.Mets=discovery_dataset_metadata$Distant_Mets, Loc.to.Met=discovery_dataset_metadata$Loc.to.Met) rownames(annotation_col)=as.vector(discovery_dataset_metadata$names) ann_colors = list( Loc.to.Met = c(Yes='blue'), Distant.Mets = c(Y='green', N='purple') ) pdf('/athena/masonlab/scratch/users/nai2008/PNET_FinnertyProject/_pdfs/discovery_analysis_heatmaps_pearsonCorrelation.pdf') pheatmap(clean_vst_counts[mm_validated_genes_w_concordantLFCs,], color=colorRampPalette(rev(brewer.pal(n = 11, name = "RdBu")))(3000), main='Discovery Dataset; Pearson correlation', clustering_distance_rows = "correlation", clustering_distance_cols = "correlation", cluster_rows=TRUE, cluster_cols=TRUE, annotation_col=annotation_col, scale='row', fontsize_col=5, annotation_colors = ann_colors) dev.off() tt=clean_vst_counts[mm_validated_genes_w_concordantLFCs,] shapiro.test(as.vector(tt[10,]))
841ec77d0fe5068d6a7613c2b6101fd64491c617
b353f7d6f9f514b889c88b29f8b03f6eb219366a
/man/clean_yahoo_data.Rd
cb5d754ae51b0f1ed1616da3827c96f563f6a765
[]
no_license
ces0491/companyDataScrapeR
806b83fa2068267bf6435e51096e4c4125642962
76ba4ddb4a42003a966cce84dff238a6e7eb89fd
refs/heads/master
2023-06-16T04:05:15.530547
2021-07-15T17:22:13
2021-07-15T17:22:13
291,532,715
1
0
null
null
null
null
UTF-8
R
false
true
427
rd
clean_yahoo_data.Rd
% Generated by roxygen2: do not edit by hand % Please edit documentation in R/clean_yahoo_data.R \name{clean_yahoo_data} \alias{clean_yahoo_data} \title{clean Yahoo data} \usage{ clean_yahoo_data(scraped_data, type, frequency = NULL) } \arguments{ \item{scraped_data}{tbl_df of scraped data} \item{type}{string indicating the data type, e.g. 'IS'} \item{frequency}{string} } \value{ tbl_df } \description{ clean Yahoo data }
365a625d8bbb741a30809cb1f4df9a3ef29aada5
e8a4396d8d60cb3d530fd977c9864c651c143efa
/analysis/additional_within_cross_models_and_plots.R
ee6eb66cb2ecf5e3a95bfa0d860ea4e6c994bbfc
[ "MIT" ]
permissive
joshsbloom/yeast-16-parents
5d8a8fb837966b9150cbb22a0dd3fd1f6bed50dd
c913c9ae7fd237329f639de02e7ec511b048730f
refs/heads/master
2020-04-17T09:34:56.206569
2019-09-23T22:15:12
2019-09-23T22:15:12
166,464,371
1
1
null
null
null
null
UTF-8
R
false
false
10,157
r
additional_within_cross_models_and_plots.R
# plot composite mapping per cross peaksModel=list() interactionPeaks=list() marginalR=list() for(cross.name in crosses) { print(cross.name) cross=cross.list[[cross.name]] if(cross.name=='A') { cross=subset(cross, ind=!grepl('A11', as.character(cross$pheno$id))) } snames = as.character(cross$pheno$id) subPheno=lapply(NORMpheno, function(x) x[match(snames, names(x))]) mPheno =sapply(subPheno, function(x) sapply(x, mean, na.rm=T)) mPheno=apply(mPheno,2, function(x) {x[is.na(x)]=mean(x, na.rm=T); return(x)}) g=pull.argmaxgeno(cross) # are there fixed loci ?? (no)------------------------------- #g.af=apply(g,2,function(x) sum(x==1)) #parents.list[[cross.name]]$fixed=(g.af==0 | g.af==nrow(g)) #fixed.loci=which(parents.list[[cross.name]]$fixed) #if(length(fixed.loci)>0) { g=g[,-fixed.loci] } #------------------------------------------------------------ #g.r=g[,-which(duplicated(g, MARGIN=2))] g.s=scale(g) #A=tcrossprod(g.s)/(ncol(g.s)) mPheno=scale(mPheno) marginalR[[cross.name]]=(crossprod(mPheno,g.s)/(nrow(mPheno)-1)) #.1581 = var exp ~.05 #.2236 = var exp ~.16 cps=cross.peaks[[cross.name]] cps=cps[cps$q<.05,] # remove 4NQO, YPD;;2 and YPD;;3 for(pheno in names(subPheno)[-c(1,38,39)]) { print(pheno) cpQTL=cps[cps$trait==pheno,] if(length(cpQTL$pmarker)!=0) { apeaks = unique(match(cpQTL$fscan.markers, colnames(g.s))) X=data.frame(g.s[,apeaks]) fitme=lm(mPheno[,pheno]~.-1, data=X) aov.a = anova(fitme) tssq = sum(aov.a[,2]) a.effs=(aov.a[1:(nrow(aov.a)-1),2]/tssq) coeffs=coefficients(fitme) cpQTL$var.exp=a.effs cpQTL$lm.coeff=as.vector(coeffs) cpQTL$chr=sapply(strsplit(cpQTL$pmarker, '_'), function(x) x[1]) cpQTL$pos=as.numeric(sapply(strsplit(cpQTL$pmarker, '_'), function(x) x[2])) cpQTL$cross=cross.name names(cpQTL)[1]='trait' print(cpQTL) peaksModel[[cross.name]][[pheno]]=cpQTL } if(length(cpQTL$pmarker)>1) { qtl.combs=combn(apeaks,2) null=lm(mPheno[,pheno]~g.s[,apeaks]-1) int.coef1=rep(NA, ncol(qtl.combs)) int.coef2=rep(NA, ncol(qtl.combs)) int.coef=rep(NA, ncol(qtl.combs)) int.pvalue=rep(NA, ncol(qtl.combs)) for(ist in 1:ncol(qtl.combs)){ full=lm(mPheno[,pheno]~g.s[,apeaks]+g.s[,qtl.combs[1,ist]]*g.s[,qtl.combs[2,ist]]-1) int.pvalue[ist]=anova(null,full)$'Pr(>F)'[2] int.coef1[ist]=coef(full)[paste0("g.s[, apeaks]",colnames(g.s)[qtl.combs[1,ist]])] int.coef2[ist]=coef(full)[paste0("g.s[, apeaks]",colnames(g.s)[qtl.combs[2,ist]])] int.coef[ist]=coef(full)[length(coef(full))] #anova(null,full)$'Pr(>F)'[2] } tqc=t(qtl.combs) dfi=data.frame(m1=colnames(g.s)[tqc[,1]], m2=colnames(g.s)[tqc[,2]], int.coef1, int.coef2, int.coef, int.pvalue, stringsAsFactors=F) dfi$cross=cross.name dfi$chr1=sapply(strsplit(dfi$m1, '_'), function(x) x[1]) dfi$chr2=sapply(strsplit(dfi$m2, '_'), function(x) x[1]) dfi$pos1=as.numeric(sapply(strsplit(dfi$m1, '_'), function(x) x[2])) dfi$pos2=as.numeric(sapply(strsplit(dfi$m2, '_'), function(x) x[2])) dfi$trait=pheno interactionPeaks[[cross.name]][[pheno]]=dfi #interactions_per_trait[[pheno]]=dfi } } } #save(marginalR,file='/data/rrv2/genotyping/RData/FDR_marignalR.RData') #save(peaksModel,file='/data/rrv2/genotyping/RData/FDR_wcPeaksModel.RData') #save(interactionPeaks, file='/data/rrv2/genotyping/RData/FDR_wcInteractionPeaksModel.RData') load('/data/rrv2/genotyping/RData/FDR_marignalR.RData') load('/data/rrv2/genotyping/RData/FDR_wcPeaksModel.RData') load('/data/rrv2/genotyping/RData/FDR_wcInteractionPeaksModel.RData') cross.peaks.flat=do.call('rbind', lapply(peaksModel, function(y) { do.call('rbind', y)} )) #cross.peaks) cross.peaks.flat$gcoord=gcoord.key[cross.peaks.flat$chr]+cross.peaks.flat$pos interactionPeaks.flat=do.call('rbind', lapply(interactionPeaks, function(y){ do.call('rbind', y)})) qs.int=qvalue(interactionPeaks.flat$int.pvalue, fdr.level=.1) interactionPeaks.flat$significant=qs.int$qvalues<.1 intP=interactionPeaks.flat[interactionPeaks.flat$significant,] intP$gcoord1=gcoord.key[intP$chr1]+intP$pos1 intP$gcoord2=gcoord.key[intP$chr2]+intP$pos2 intP=na.omit(intP) ssi=split(intP, paste(intP$trait, intP$cross) ) hist(c(sapply(ssi, nrow), rep(0, (38*16)-length(ssi)))) glength=sum(unlist(chr.lengths)) #load('/data/rrv2/genotyping/RData/jointPeaks5.RData') #jP=rbindlist(jointPeaks5, idcol='chromosome') #jPs=split(jP, jP$trait) jointPeaksFlat=rbindlist(jointPeaksJS, idcol='chromosome') #data.frame(do.call('rbind', jointPeaks5), stringsAsFactors=F) names(jointPeaksFlat)[1]='chr' #jointPeaksFlat$chr=sapply(strsplit(jointPeaksFlat$marker, '_'), function(x) x[1]) jointPeaksFlat$pos=as.numeric(sapply(strsplit(jointPeaksFlat$fscan.markers, '_'), function(x) x[2])) jointPeaksFlat$gpos=gcoord.key[jointPeaksFlat$chr ]+jointPeaksFlat$pos utraits.orig=unique(cross.peaks.flat$trait) utraits=utraits.orig utraits[34]='YNB_ph8' utraits[36]='YPD_15C' utraits[33]='YNB_ph3' utraits[10]='EtOH_Glu' utraits[37]='YPD_37C' utraits=gsub(';.*','', utraits) utraits=gsub('_', ' ', utraits) pdf(file=paste0('/home/jbloom/Dropbox/RR/Figures and Tables/SupplementaryFigure2.pdf'), width=11, height=8) for(piter in 1:length(utraits)) { png(file=paste0('/home/jbloom/Dropbox/RR/Figures and Tables/other formats/SuplementaryFigure2_', piter, '.png'), width=1100, height=800) pheno.iter=utraits.orig[piter] #pdf(file=paste0('/home/jbloom/Dropbox/RR/Figures and Tables/', filename.clean(pheno.iter), '_joint.pdf'), width=1024, height=800) pcnt=0 op <- par(mfrow = c(16,1), oma = c(5,8,5,.5) + 0.01, mar = c(0,4,0,0) + 0.01, xaxs='i', yaxs='i' ) joint.peaks.toplot=jointPeaksFlat[jointPeaksFlat$trait==pheno.iter,] joint.peaks.toplot=joint.peaks.toplot[!duplicated(joint.peaks.toplot$gpos),] parent.vec=c('M22', 'BY', 'RM', 'YPS163', 'YJM145', 'CLIB413', 'YJM978', 'YJM454', 'YPS1009', 'I14', 'Y10', 'PW5', '273614N', 'YJM981', 'CBS2888', 'CLIB219') glength=1.2e7 for(cross.iter in 1:length(crosses)){ cross.name=crosses[cross.iter] jptlookup=joint.peaks.toplot[joint.peaks.toplot$fscan.markers %in% parents.list[[cross.name]]$marker.name,] cross.sub.p=cross.peaks.flat[cross.peaks.flat$trait==pheno.iter & cross.peaks.flat$cross==crosses[cross.iter],] cross.sub.pi=intP[intP$trait==pheno.iter & intP$cross==crosses[cross.iter],] mpM=marginalR[[crosses[cross.iter]]] mpM.marker=tstrsplit(colnames(mpM), '_', type.convert=T) mpM.gcoord=gcoord.key[mpM.marker[[1]]]+mpM.marker[[2]] if(nrow(cross.sub.p)>0) { #plot(0,0, type='n', xlim=c(0, glength), yaxt='n', ylab='', xaxt='n', yaxt='n', ylim=c(-1,1),cex.lab=1.5) plot(0,0, type='n', xlim=c(0, glength), yaxt='n', ylab='', xaxt='n', yaxt='n', ylim=c(0,1),cex.lab=1.5) abline(h=0) abline(v=jptlookup$gpos, col='lightgreen') axis(2,at=1,labels=parent.vec[cross.iter], cex.axis=1.5, las=2) pcnt=pcnt+nrow(cross.sub.p) signme=sign(cross.sub.p$lm.coeff) if(min(grep(crosses.to.parents[[cross.iter]][1], names(parents.list[[cross.iter]])))==7) { signme=signme signR=-1*mpM[pheno.iter,] } else{ signme=-1*signme signR=mpM[pheno.iter,] } # flip sign ... (if negative then point to strain that increases growth) signme=-1*signme splus=signme==1 sminus=signme==-1 #points(mpM.gcoord, signR) cross.sub.p$lm.ceiling=rep(.1, nrow(cross.sub.p)) cross.sub.p$lm.ceiling[cross.sub.p$var.exp>0]=.2 cross.sub.p$lm.ceiling[cross.sub.p$var.exp>.04]=.5 cross.sub.p$lm.ceiling[cross.sub.p$var.exp>.08]=.75 cross.sub.p$lm.ceiling[cross.sub.p$var.exp>.25]=1 if(sum(splus)>0){ arrows(cross.sub.p$gcoord[splus],0, cross.sub.p$gcoord[splus], cross.sub.p$lm.ceiling[splus], code=2, length=.12, lwd=4, col=ifelse(cross.sub.p$lm.ceiling[splus]>.2, 'black', 'grey')) } if(sum(sminus)>0){ arrows(cross.sub.p$gcoord[sminus],cross.sub.p$lm.ceiling[sminus], cross.sub.p$gcoord[sminus],0 , code=2, length=.12, lwd=4, col=ifelse(cross.sub.p$lm.ceiling[sminus]>.2, 'black', 'grey') ) } abline(v=gcoord.key, lty=2, col='lightblue') } else { plot(0,0, type='n', xlim=c(0, max(glength)), ylim=c(0,1), xaxt='n' ) #ylab=crosses[cross.iter] , #abline(h=0, lty=3, col='grey') abline(v=cumsum(genome.chr.lengths), lty=2, col='lightblue') } # if(nrow(cross.sub.pi)>0) { # peak.number=c(seq_along(cross.sub.pi[,1]), c(seq_along(cross.sub.pi[,2]))) # #peak.chr=c(cross.sub.pi$chr1, cross.sub.pi$chr2) # #peak.pos=as.numeric(sapply(strsplit(sapply(strsplit(c(cross.sub.pi.sig[,1], cross.sub.pi.sig[,2]), ':'), function (x) x[2]), '_'), function(x)x[1])) # peak.gpos=c(cross.sub.pi$gcoord1, cross.sub.pi$gcoord2) # text(peak.gpos, (peak.number/max(peak.number))*.9, '*', col='red', cex=4) # } if(cross.iter==16){ axis(1, at=gcoord.key, labels=names(gcoord.key), cex.axis=1.5)} } title(xlab='genomic position', ylab='', outer=TRUE, cex.lab=2, main=paste(utraits[piter], ' ', pcnt, 'total QTL | ', length(joint.peaks.toplot$gpos), 'joint QTL' )) dev.off() } # dev.off()
7f060fe58c8d3475e040e55373691ab2d39b40b2
a3da395d683014c2f04a4491f5cf3214076a82f6
/fmd work.R
6fc42b194ebf9d748e93a5907a75f8e8b1811207
[]
no_license
VetMomen/UBM
f2f9ba780a8ce09aba3a2c64c29ba06d93e49a37
d4af7f433144b4f7bf62335ee8e4ad6bb1997495
refs/heads/master
2020-04-02T02:25:25.759831
2019-03-30T13:44:54
2019-03-30T13:44:54
151,887,463
0
0
null
null
null
null
UTF-8
R
false
false
4,352
r
fmd work.R
dir.create('./plots') fmd<-read_excel(path = './data sets/FMD form (Responses).xlsx',col_types = c('guess','guess','guess','guess','guess','numeric','numeric','guess','guess','guess','guess','guess','guess')) str(fmd) uniqus<-apply(fmd,2,unique) #frequancy of Each type and location xtab<-with(fmd, table(`Production Type`,Location)) par(las=1) plot(xtab,main='distribution of production type and location') #total cap in Eche loc cap_loc<-fmd%>%group_by(Location)%>%summarize(Cap=sum(Cap.)) cap_loc%>%ggplot(aes(Location,Cap))+ geom_col(aes(fill=Location),width = .7)+ geom_text(data = cap_loc,aes(label=Cap),vjust=.003,color='darkseagreen2')+ ylab('Capacity')+ ggtitle('Capacity of each area')+ theme(panel.background = element_rect(fill = 'black'),panel.grid.minor = element_line(colour = 'black')) #total cap of each type cap_type<-fmd%>%group_by(`Production Type`)%>%summarize(Cap=sum(Cap.)) cap_type%>%ggplot(aes(`Production Type`,Cap,fill=`Production Type`))+ geom_col(width = .7)+ geom_text(data = cap_type,aes(label=Cap),vjust=.003,color='darkseagreen2')+ ylab('Capacity')+ ggtitle('Capacity of each type')+ theme(panel.background = element_rect(fill = 'black'),panel.grid.minor = element_line(colour = 'black')) #mixing area with type cap_type<-fmd%>%group_by(`Production Type`,Location)%>%summarize(Cap=sum(Cap.)) cap_type%>%ggplot(aes(`Production Type`,Cap,fill=`Production Type`))+ geom_col(width = .7)+ geom_text(data = cap_type,aes(label=Cap),vjust=.003,color='darkseagreen2')+ ylab('Capacity')+ ggtitle('Capacity of each type in each area')+ theme(panel.background = element_rect(fill = 'black'),panel.grid.minor = element_line(colour = 'black'))+ facet_wrap(.~Location) #adding factor of infection cap_type<-fmd%>%group_by(`Production Type`,Location,infected)%>%summarize(Cap=sum(Cap.)) cap_type%>%ggplot(aes(`Production Type`,Cap,fill=infected))+ geom_col(width = .7)+ geom_text(data = cap_type,aes(label=Cap),vjust=.003,color='darkseagreen2')+ ylab('Capacity')+ ggtitle('Capacity of each type in each area illustrating infected herd')+ theme(panel.background = element_rect(fill = 'black'),panel.grid.minor = element_line(colour = 'black'))+ facet_grid(.~Location) #adding vaccine type cap_type<-fmd%>%group_by(`Production Type`,Location,infected,`vacc. Type`)%>%summarize(Cap=sum(Cap.)) cap_type%>%ggplot(aes(`Production Type`,Cap,fill=infected))+ geom_col(width = .7)+ geom_text(data = cap_type,aes(label=Cap),vjust=.003,color='darkseagreen2')+ ylab('Capacity')+ ggtitle('Capacity of each area illustrating infected herd & vaccine type')+ theme(panel.background = element_rect(fill = 'black'),panel.grid.minor = element_line(colour = 'black'))+ facet_grid(`vacc. Type`~Location) #farm location color<-colorFactor(palette =c('blue','red') ,domain = fmd$infected) fmd%>%leaflet()%>% addProviderTiles(providers$OpenStreetMap.BlackAndWhite)%>%addCircleMarkers(lat = fmd$lat, lng = fmd$lon,color = ~color(fmd$infected), radius = fmd$Cap./1000)%>% addLegend(position = 'topright',pal = color,values = ~factor(fmd$infected),title = 'Infection') perc<-fmd%>%group_by(`vacc. Type`)%>%summarize(percent=(sum(Cap.)/sum(fmd$Cap.))*100) perc$percent<-round(perc$percent,1) fmd%>%group_by(`vacc. Type`)%>%summarize(total=sum(Cap.), percent=round((sum(Cap.)/sum(fmd$Cap.))*100,1))%>% ggplot(aes(x='',y=total,fill=`vacc. Type`))+ geom_col(width = .3)+ coord_polar(theta = 'y',start = 0,direction = 1,clip = 'on')+ theme(axis.title = element_text(face = 'bold'),axis.line = element_blank(), panel.background = element_blank(), axis.text = element_blank(),panel.grid = element_blank(), axis.title.y.left = element_blank(), axis.title.x.bottom = element_blank())+ geom_text(aes(label=percent),nudge_x = .2,hjust=.5)+ scale_fill_brewer(type = qualitative,palette = 'Dark2')+ labs(title = 'Vaccination type share')+ theme(plot.title = element_text(hjust = .5))
03976a400225f076749496f873d27f14b6b6e6d5
0618bcede3e643e0ce2c9b4f6ea70bceb2778b78
/rscripts/coveragePeaks.R
5d03973d62af23a8e3f59025b0857d551ffcdced
[]
no_license
emdann/HexamerBias
2f3c523b374931804c12b9a85a11c68a666f007f
b8e0065d0331efcf4a0e0337277f13b6d6ee8b55
refs/heads/master
2021-09-21T21:03:55.383020
2018-08-31T12:29:34
2018-08-31T12:29:34
null
0
0
null
null
null
null
UTF-8
R
false
false
644
r
coveragePeaks.R
## COVERAGE PEAKS library(data.table) library(dplyr) library(ggplot2) peakAnn <- fread("mnt/edann/coverage_peaks/multipeaks.annotatePeaks.homer") randAnn <- fread("mnt/edann/coverage_peaks/multipeaks.random.annotatePeaks.homer") pdf("AvOwork/output/covPeaks_distTSS_boxplot.pdf") boxplot(abs(peakAnn$`Distance to TSS`), abs(randAnn$`Distance to TSS`), outline = FALSE, varwidth = TRUE, names = c("Coverage peaks", "random"), ylab='Distance to TSS') dev.off() g <- randAnn %>% mutate(Annotation=gsub(pattern = "\\(.+\\)",replacement = "", x=Annotation)) %>% ggplot(., aes(Annotation)) + geom_bar() randAnn %>% ggplot(., aes(Annotation))
d9ab4ba748a72ba899a8d5da2b5dd422e3e83632
ac644d6d019e20b8a54df53ad99959ea641c1d49
/sentimentAnalysis.R
af487847bbc07653828d3d42cd6716eb84e75245
[]
no_license
Nithya945/Crime-Trends-in-the-City
5d65a29635234a47835561138503c2039af7ec37
87abc4bf67e80161bf1eb3710d4566cb9c3ddda9
refs/heads/master
2021-05-11T20:13:35.229892
2018-04-11T07:53:11
2018-04-11T07:53:11
117,435,078
0
0
null
null
null
null
UTF-8
R
false
false
8,451
r
sentimentAnalysis.R
library(maptools) library(plyr) library(ggplot2) library(car) library(MASS) library(sp) require("RPostgreSQL") source("/home/nithya/Desktop/Crime-Prediction-master/CrimePredictionUtil.R") #################################### ## Twitter data prerprocessing ##### #################################### # build the link to my PostgreSQL database drv <- dbDriver("PostgreSQL") print(drv) con <- dbConnect(drv, host = 'localhost', port='5432', dbname = 'postgres',user = 'postgres', password = 'apple945') # draw query from PostgresSQL database # get tweets from "2014-01-01 00:00:00" to "2014-01-31 11:59:59" Jan.2014 <- tweet.qry("2014-01-01 00:00:00", "2014-01-31 11:59:59") print(Jan.2014) # read chicago boundary city.boundary = read.shapefile("/home/nithya/Desktop/Crime-Prediction-master/City_Boundary/City_Boundary/City_Boundary.shp", "poly", "+init=epsg:3435", "+init=epsg:26971") city.boundary # set up grid (neighborhood) for concatenating tweets bb <- bbox(city.boundary) # bbox of city boundary bb cs <- c(1000, 1000) # cell size 1000m *1000m cs cc <- bb[, 1] + (cs/2) # cell offset cc cd <- ceiling(diff(t(bb))/cs) # number of cells per direction cd grid <- GridTopology(cellcentre.offset=cc, cellsize=cs, cells.dim=cd) # create a grib topology grid data=data.frame(id=1:prod(cd)) data proj4string=CRS(proj4string(city.boundary)) proj4string # conver grid topology to spatial data frame sp_grid <- SpatialGridDataFrame(grid, data=data.frame(id=1:prod(cd)), proj4string=CRS(proj4string(city.boundary))) class(sp_grid) head(sp_grid@data,100) str(sp_grid) summary(sp_grid) plot(city.boundary, xlim=c(332777, 367345), ylim=c(552875, 594870)) plot(sp_grid, add =TRUE) spplot(sp_grid, sp.layout = c("sp.points", SpatialPoints(coordinates(sp_grid)))) Jan.2014 # convert xy coordinate of tweets as spatial points Jan.2014.xy <- Jan.2014[7:8] Jan.2014.xy coordinates(Jan.2014.xy) <- ~ st_x+st_y proj4string(Jan.2014.xy) <- proj4string(city.boundary) points(Jan.2014.xy, pch =".") Jan.2014.xy # assign tweets to each grid (neighbourhood) Jan.tweet.grid <- over(Jan.2014.xy, sp_grid) Jan.tweet.grid names(Jan.tweet.grid) <- "grid_id" Jan.2014.grid <- cbind(Jan.2014, Jan.tweet.grid) Jan.2014.grid # convert date type Jan.2014.grid$dates <- as.POSIXct(Jan.2014.grid$date) Jan.2014.grid # splite time period into every 6-hour dates.combine.tweets.crime <- c(Jan.2014.grid$dates, theft.2014.jan.to.feb[which(theft.2014.jan.to.feb$month==1),]$timestamp) dates.combine.tweets.crime ncol(dates.combine.tweets.crime) factor.combine <- cut(dates.combine.tweets.crime, "6 hours") length(factor.combine) factor.combine Jan.2014.grid$sixhr <- factor.combine[1:length(Jan.2014.grid$dates)] # levels(Jan.2014.grid$sixhr)[1:10] Jan.2014.grid$sixhr_n <- as.numeric(Jan.2014.grid$sixhr) Jan.2014.grid$dates <- NULL #names(Jan.test.grid$date) #class(as.POSIXlt(Jan.test.grid$created_at[1:5])) # concatenate tweets with same grid_id and 6-hour period together Jan.2014.paste <- ddply(Jan.2014.grid, c("sixhr_n", "grid_id"), summarise, text_p=paste(tweet, collapse=" ")) text print(Jan.2014.paste) # clean-up twitter data using twitter.clean function Jan.2014.paste.c <- twitter.clean(Jan.2014.paste, Jan.2014.paste$text_p) Jan.2014.paste.c row.to.keep <- !is.na(Jan.2014.paste.c$grid_id) Jan.2014.paste.c <- Jan.2014.paste.c[row.to.keep,] print(Jan.2014.paste.c) #summary(Jan.2014.paste.c) #save("Jan.2014.paste.c", file = "Capstone/Jan_2014_paste_c_6hr.Rdata") #load("Capstone/Jan_2014_paste_c_6hr.Rdata") ################################### ### calculate snetiment score ##### ################################### Jan.2014.pol.6h <- NULL #options(warn=0, error = recover) # load polarity file created by Lexicon.R load("/home/nithya/POLKEY.RData") # calculate sentiment score system.time( Jan.2014.pol.6h<- polarity(Jan.2014.paste.c$text1, grouping.var = NULL, polarity.frame = POLKEY, constrain = TRUE, negators = qdapDictionaries::negation.words, amplifiers = qdapDictionaries::amplification.words, deamplifiers = qdapDictionaries::deamplification.words, question.weight = 0, amplifier.weight = .3, n.before = 4, n.after = 2, rm.incomplete = FALSE, digits = 3) ) # user system elapsed # 5777.63 1.81 5841.16 # save("Jan.2014.pol.6h",file = "Jan_2014_pol_1000m_6h.Rdata") # load("Jan_2014_pol_1000m_6h.Rdata") # str(Jan.2014.pol.6h) # Jan.2013.paste.c$text1[1] # Jan.2014.paste.c$pol <- Jan.2014.pol$all$polarity # test <- ddply(Jan.2014.paste.c, c("sixhr"), summarise, # difference = diff(pol,2)) # # Jan.2014.combined <- cbind(Jan.2014.paste.c[63:33498,], test[,2]) # test2 <- Jan.2014.combined[which(Jan.2014.combined$grid_id == 587),] # head(test2) # plot(test2[,6]/10) # calculate mean from raw score Jan.2014.pol.6h$mean <- mean(Jan.2014.pol.6h$all$polarity) Jan.2014.pol.6h # center the data by subtracting $sum from $mean Jan.2014.pol.6h$all$centered <- Jan.2014.pol.6h$all$polarity - Jan.2014.pol.6h$mean # plot sentiment score without centering qplot(Jan.2014.pol.6h$all$polarity, main = "Sentiment Histogram", xlab = "Score", ylab = "Frequency", binwidth = 0.015) # plot centered sentiment score qplot(Jan.2014.pol.6h$all$centered, main = "Centered Sentiment Histogram", xlab = "Score", ylab = "Frequency", binwidth = 0.075) # insert day of month and grid_id into large polarity Jan.2014.pol.6h$all$sixhr_n <- Jan.2014.paste.c$sixhr_n + 1 #shift a 6-hour period Jan.2014.pol.6h$all Jan.2014.pol.6h$all$grid_id <- Jan.2014.paste.c$grid_id Jan.2014.pol.6h$all # summary(Jan.2014.pol.6h$all$grid_id) # create data.frame which contains 6-hour period, polarity and grid_id Jan.2014.pol.6h.data <- data.frame() Jan.2014.pol.6h.data <- data.frame(Jan.2014.pol.6h$all$sixhr, Jan.2014.pol.6h$all$grid_id, Jan.2014.pol.6h$all$polarity) names(Jan.2014.pol.6h.data) <- c("sixhr_n", "grid_id", "polarity") Jan.2014.pol.6h.data # inset missing row from ddply vals <- expand.grid(sixhr_n = 2:123,grid_id = 1:max(Jan.2014.pol.6h.data$grid_id, na.rm = TRUE)) head(vals) summary(vals) Jan.2014.pol.6h.data.m <- merge(vals, Jan.2014.pol.6h.data,all.x=TRUE) Jan.2014.pol.6h.data.m Jan.2014.pol.6h.data.m[which(Jan.2014.pol.6h.data.m$polarity!='NA'),] #summary(Jan.2014.pol.6h.data.m) # impute 0 to those missing polarity Jan.2014.pol.6h.data.m[is.na(Jan.2014.pol.6h.data.m$polarity),"polarity"] <- 0 # calculate trend index for all grid area Jan.2014.pol.trend.6hour1 <- data.frame() Jan.2014.pol.trend.6hour <- Jan.2014.pol.6h.data.m[which(Jan.2014.pol.6h.data.m$polarity!='NA'),] Jan.2014.pol.trend.6hour system.time( for (i in 1:max(Jan.2014.pol.6h.data$grid_id, na.rm = TRUE)){ sub <- subset(Jan.2014.pol.trend.6hour, grid_id == i) sub sub$trend_3 <- trend.idx(sub$polarity,3,0.1) Jan.2014.pol.trend.6hour1 <- rbind(Jan.2014.pol.trend.6hour1,sub) } ) str(Jan.2014.pol.trend.6hour) summary(Jan.2014.pol.trend.6hour) #save(Jan.2014.pol.trend.6hour, file = "Capstone/allsub_6hr.Rdata") #View(Jan.2014.pol.trend.6hour) # visualize sentiment score and its trend in neighbourhood 587 (downtown sub.587 <- subset(Jan.2014.pol.6h$all, grid_id == 587)[,c(3,7)] # bb.scatter.587 bb.scatter.587 <- ggplot(sub.587, aes(x = sub.587$sixhr_n, y = sub.587$polarity)) bb.scatter.587 <- bb.scatter.587 + geom_point() + geom_line() + ylim(-1, 1) bb.scatter.587 <- bb.scatter.587 + xlab("Period") + ylab("Sentiment") + ggtitle("Neighborhood 587") bb.scatter.587 # calsulate trend index sub.587$trend_2 <- trend.idx(sub.587$polarity,2,0.1) sub.587$trend_3 <- trend.idx(sub.587$polarity,3,0.1) # plot trend index for each 12-hour t_2.scatter.587 <- ggplot(sub.587, aes(x = sub.587$mday, y = sub.587$trend_2)) t_2.scatter.587 <- t_2.scatter.587 + geom_point() + geom_line() t_2.scatter.587 <- t_2.scatter.587 + xlab("Date") + ylab("trend_2") + ggtitle("587") t_2.scatter.587 # plot trend index for each 18-hour t_3.scatter.587 <- ggplot(sub.587, aes(x = sub.587$sixhr_n, y = sub.587$trend_3)) t_3.scatter.587 <- t_3.scatter.587 + geom_point() + geom_line() t_3.scatter.587 <- t_3.scatter.587 + xlab("Period") + ylab("Trend_3") + ggtitle("Neighborhood 587") t_3.scatter.587 # use multiplot function to plot both trend index multiplot(bb.scatter.587, t_3.scatter.587)
fdcaf7e68084750902372aeab4cda547bb61babd
d743759238db4f45f2ceef30118008e52a67ea63
/RCode/Data/0_cleanWTO.R
56cc5a2deda7ddfd0dbf3eb8875e9dcff71898fc
[]
no_license
cindyyawencheng/Codex
e628ace6d070c46b6c7896ff01468e39603fdfbf
30e178c7fc8e7d295512c3c7f434c175a69d6781
refs/heads/master
2020-04-06T23:39:44.345165
2019-09-26T13:51:09
2019-09-26T13:51:09
157,876,928
0
0
null
null
null
null
UTF-8
R
false
false
2,605
r
0_cleanWTO.R
# clean WTO membership data if(Sys.info()['user'] == 'cindycheng'){ source('/Users/cindycheng/Documents/Papers/Codex/RCode/setup.R') } # ------------------------- # Clean WTO Membership data # -------------------------- wto = read.csv(paste0(pathMain, '/participation_development/mem-obs-list.csv'), stringsAsFactors = FALSE) wto$date = as.Date(wto$Membership.Date, '%d %B %Y') wto$year = as.numeric(format(wto$date, "%Y")) # observers = wto[166:188,] # observers$Members[observers$Members == 'Iran'] = "Iran (Islamic Republic of)" # observers$Members[observers$Members == "Lebanese Republic"] = "Lebanon" # observers$Members[observers$Members == "Sao Tome and Principe"] = "Sao Tomé and Principe" # observers$Members[observers$Members == "Sudan"] = "Sudan (North + South)" # # observers$observerDum = 1 wto = wto[1:164,] wto$wtoDum = 1 wto$Members[wto$Members == "Bahrain, Kingdom of" ] = 'Bahrain' wto$Members[wto$Members == "Bolivia, Plurinational State of" ] = "Bolivia" wto$Members[wto$Members == "Côte d’Ivoire" ] = "Côte d'Ivoire" wto$Members[wto$Members == "European Union (formerly EC)"] = "European Union" wto$Members[wto$Members == "Kuwait, the State of"] = "Kuwait" wto$Members[wto$Members == "Kyrgyz Republic"] = "Kyrgyzstan" wto$Members[wto$Members == "Lao People’s Democratic Republic"] = "Lao People's Democratic Republic" wto$Members[wto$Members == "Moldova, Republic of"] = "Republic of Moldova" wto$Members[wto$Members == "Saudi Arabia, Kingdom of"] = "Saudi Arabia" wto$Members[wto$Members == "Slovak Republic"] = "Slovakia" wto$Members[wto$Members == "Chinese Taipei"] ="Taipei, Chinese" wto$Members[wto$Members == "Tanzania"] = "Tanzania, United Republic of" wto$Members[wto$Members == "United States"] = "United States of America" wto$Members[wto$Members == "Venezuela, Bolivarian Republic of"] = "Venezuela" wto$Members[wto$Members == "North Macedonia"] = "Macedonia, The Former Yugoslav Republic of" particip[which(particip$actor_name == 'European Union' & particip$event_short == 'CCEURO' & particip$year == "1996" ),] particip[which(part)] wtoLong = expand.grid(wto$Members %>% unique() %>% sort(), 1964:2018) names(wtoLong) = c('Members', 'year') wtoLong = wtoLong[order(wtoLong$Members, wtoLong$year),] wtoLong$wtoDum = wto$wtoDum[match(paste0(wtoLong$Members, wtoLong$year), paste0(wto$Members, wto$year))] wtoLong$wtoDum[which(is.na(wtoLong$wtoDum ))] = 0 wtoLong$wtoDum = unlist(lapply(split(wtoLong$wtoDum, wtoLong$Members), cumsum)) save(wtoLong, file = paste0(pathMain, '/participation_development/mem-obs-list_wtoclean.rda'))
875e4cd3a6026caae37284e75dcb23328798676b
6e2547b89094ab73aba4b39fb79270e317d1a001
/monte_carlo_triangle.r
9751b3eb2dc0772700d8e5767b9f044c8cd43228
[]
no_license
kjlockhart/RAPfish
bf3cbb81936a4eab42e2ab886b8209d3e7cedcd6
cfb5a5ce98d626c29b9482e41371b9984e3150b9
refs/heads/master
2021-01-01T18:34:39.650166
2012-09-09T23:46:49
2012-09-09T23:46:49
5,730,374
2
0
null
null
null
null
UTF-8
R
false
false
3,370
r
monte_carlo_triangle.r
#Triangular Distribution Monte Carlo for Rapfish # 2012-04-15 Divya Varkey Created #to be obtained from site nsim=100 num_fish=53 anchor_files=c('anchors4.csv','anchors5.csv','anchors6.csv','anchors7.csv','anchors8.csv','anchors9.csv','anchors10.csv','anchors11.csv','anchors12.csv') filenames=c('CCRF_Field1.csv','CCRF_Field2.csv','CCRF_Field3.csv','CCRF_Field4.csv','CCRF_Field5.csv','CCRF_Field6.csv') ###########generated here nfield=length(filenames) discipline.names =strsplit(filenames, ".csv") images_tri=paste(discipline.names,"_Triangle_MC",".jpg",sep="") res_tri=paste("MC_Triangle_",discipline.names,".csv",sep="") source("rapfish_functions.R") source("functiontri.r") L1=num_fish+1 L2=L1+num_fish-1 U1=L2+1 U2=U1+num_fish-1 ####################TRIANGULAR MC Trig_RapfishMC<-function(fisheries.all,num_fish,nsim) { n_att=ncol(fisheries.all) fisheries.dat=fisheries.all[1:num_fish,] lb=fisheries.all[L1:L2,] ub=fisheries.all[U1:U2,] anchors=loaddata(anchor_files[n_att-3]) colnames(anchors)<-colnames(fisheries.all) mc_init=array(data=0,dim=c(num_fish,n_att,nsim)) for(j in 1:num_fish) { for(k in 1:n_att) { mc_init[j,k,]=rtri(nsim,lb[j,k],ub[j,k],fisheries.dat[j,k]) } } fish_mc_res=array(data=0,dim=c(num_fish+nrow(anchors),2,nsim)) for(m in 1:nsim) { fish_mc.dat=round(mc_init[1:num_fish,1:n_att,m],1) colnames(fish_mc.dat)<-colnames(fisheries.all) fish_mc.raw=rbind(anchors,fish_mc.dat) fish_mc.scaled = mdscale(fish_mc.raw) fish_mc_res[,,m]=fish_mc.scaled } output<-list() output$mc_init=mc_init output$fish_mc_res=fish_mc_res return(output) } ###############################MC PLOTS for(i in 1:nfield) { fisheries.all = loaddata(filenames[i]) tt=Trig_RapfishMC(fisheries.all,num_fish,nsim) n_att=ncol(fisheries.all) fisheries.dat=fisheries.all[1:num_fish,] anchors=loaddata(anchor_files[n_att-3]) colnames(anchors)<-colnames(fisheries.all) fisheries.raw=rbind(anchors,fisheries.dat) fisheries.scaled = mdscale(fisheries.raw) n_an=nrow(anchors) plot1=n_an+1 plot2=n_an+num_fish cols=rainbow(num_fish,start=0, end=.7) jpeg(filename=images_tri[i],width=20,height=16,units="cm",res=500) Res=ifelse(nfield>30,RAPplot1(fisheries.scaled,num_fish,n_an),RAPplot2(fisheries.scaled,num_fish,n_an)) mtext(side=3, line=1, "Triangular MC",adj=1) mtext(side=3, line=1, discipline.names[i],adj=0) for(m in 1:nsim) { mcplot=tt$fish_mc_res[plot1:plot2,,m] mcplot = mcplot[order(fisheries.scaled[plot1:plot2,1]),] points(mcplot,xlab="",ylab="",col=cols,pch='.') } dev.off() mc_summ=matrix(data=0,nrow=num_fish,ncol=12) s_mcres=tt$fish_mc_res[plot1:plot2,,] for(fs in 1:num_fish) { xx=round(quantile(s_mcres[fs,1,],probs=c(0.5,0.25,0.75,0.025,0.975)),4) yy=round(quantile(s_mcres[fs,2,],probs=c(0.5,0.25,0.75,0.025,0.975)),4) mc_summ[fs,2:6]=xx mc_summ[fs,8:12]=yy } mc_summ[,1]=round(fisheries.scaled[plot1:plot2,1],4) mc_summ[,5]=round(fisheries.scaled[plot1:plot2,2],4) colnames(mc_summ)<-c("X_Scores","Median","25%","75%","2.5%","97.5%","Y_Scores","Median","25%","75%","2.5%","97.5%") rownames(mc_summ)<-rownames(fisheries.dat) write.csv(mc_summ,res_tri[i]) }
04c1558ea820f5c27f3db676871be004005d85bb
ffdea92d4315e4363dd4ae673a1a6adf82a761b5
/data/genthat_extracted_code/rscala/tests/testthat.R
bb7f800c6e94d2efc0c4de68e19c27a4d08ab541
[]
no_license
surayaaramli/typeRrh
d257ac8905c49123f4ccd4e377ee3dfc84d1636c
66e6996f31961bc8b9aafe1a6a6098327b66bf71
refs/heads/master
2023-05-05T04:05:31.617869
2019-04-25T22:10:06
2019-04-25T22:10:06
null
0
0
null
null
null
null
UTF-8
R
false
false
56
r
testthat.R
library(testthat) library(rscala) test_check("rscala")
b0aa2a7c93c0e0782ae5da4389f1fca1a212a226
0da7e603de44ac6eb0a23381aa194b22563b02d0
/03_Analysis/PopGenNet20210915.R
ec11f784f7b2b89f1b2aea0381cad44191091927
[]
no_license
efronhofer/PopGenNet
5eddea7815afdb7f870b477ef563be57aef6e906
6f83abcdbd301b8a1c8efc0766338c252e270226
refs/heads/master
2023-04-07T20:51:35.080945
2021-09-29T06:04:46
2021-09-29T06:04:46
321,404,193
0
0
null
null
null
null
UTF-8
R
false
false
169,930
r
PopGenNet20210915.R
# PopGenNet analysis # 2021-09-15 # Author: Roman Alther, Eawag, Duebendorf, Switzerland # Works in R ver. 3.6.1 and 4.0.3 (tested on GNU/Linux, MacOS, Windows) with RStudio ver. 1.3.1093 #*#*#*#*#*#*#*#*#*#*#*#*#*#*#*#*#*#*#*#*#*#*#*#*#*#*#*#*#*#*#*#*#*#*#*#*#*#*#*# #### PREPARATION -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ #### rm(list=ls()) # start from scratch existing_data=T # Should the analysis run with pre-collated and attached data from the authors? Otherwise the required raw data need to be prepared using the script 'PopGenNet_Prep_20201210.R', creating an output folder in '02_Data_prep'. internal=F # defaults to FALSE, but was set to TRUE for publication preparation (lab internal use) log_trans=T # log-transform some explanatory variables ("betw_undir","betw_dir","degree","catch") critval <- 1.96 ## approx 95% confidence interval (plus/minus x times SD) fst=T # prepares figures and analyses for Fst as well # setwd(dirname(rstudioapi::getActiveDocumentContext()$path)) WD <- getwd() # save current directory (should correspond to 03_Analysis, source script from there) #* Load data #### setwd("..") DF <- getwd() if (internal){ prep_folder <- "Output20210913" rdata <- "PopGenNet_data" }else{ prep_folder <- "Output" rdata <- "PopGenNet_data" } load(paste0("02_Data_prep/",rdata,".RData")) setwd(WD) if (internal){ load("Gfos_data_20200925.RData") } output <- format(Sys.time(), "%Y%m%d") ##* Packages #### # Check if custom-made packages OpenSwissRiverPlot and MultiPanel are already installed, else install from SWITCHdrive if (!"OpenSwissRiverPlot" %in% installed.packages()[,"Package"]){ ulib <- NULL # option to define path for user defined library -> change NULL to desired path source("https://drive.switch.ch/index.php/s/kgoAUIbqxYc92YP/download") # install OpenSwissRiverPlot # alternatively you may install from the included .tgz file # install.packages("OpenSwissRiverPlot.tgz", repos = NULL, type="source", lib=ulib, INSTALL_opts="--no-multiarch") } if (!"MultiPanel" %in% installed.packages()[,"Package"]){ ulib <- NULL # option to define path for user defined library -> change NULL to desired path source("https://drive.switch.ch/index.php/s/tdaTpPUM7rH1P4X/download") # install MultiPanel # alternatively you may install from the included .tgz file # install.packages("MultiPanel.tgz", repos = NULL, type="source", lib=ulib, INSTALL_opts="--no-multiarch") } # Load CRAN packages library(igraph) # to do network "stuff", works with version 1.2.4.2 # library(jtools) # for "easy" model output (function summ()), works with version 2.1.1 # library(rsq) # for R-squared (function rsq()), works with version 2.0 # library(vegan) # for Mantel tests (FST by instream distance), works with version 2.5-6 # library(MuMIn) # for model selection using dredge(), works with version 1.43.17 # Load custom packages and check for updates if (internal){ # Load lab internal package for publication figure preparation library(SwissRiverPlot) # to plot maps of Switzerland, works with version 0.4-2 update_SRP() }else{ library(OpenSwissRiverPlot) # to plot maps of Switzerland, works with version 0.4-0 update_OSRP() } library(MultiPanel) # to plot multipanel figures, works with version 0.6-3 update_MultiPanel() #* Figures preparation #### ##*** Figure format #### pdf=F # set to TRUE if figures should be prepared as PDF ##*** Size #### fig.width=12 # standard figure width in inches fig.height=9 # standard figure height in inches ##*** Labels #### label_A <- expression(italic(G.~fossarum)~"type A") # italic G. fossarum A label_B <- expression(italic(G.~fossarum)~"type B") # italic G. fossarum B label_mod <- "Simulation data" D_label <- c("0.001","0.01","0.1") W_label <- c("0","0.5","1") K_label <- c("0","1") D <- paste0("disp_",D_label) W <- paste0("w_up_",W_label) K <- paste0("K_scale_",K_label) dlab <- c(as.expression(bquote(italic(d)~"="~.(D_label[[1]]))), # d = 000.1 as.expression(bquote(italic(d)~"="~.(D_label[[2]]))), # d = 00.1 as.expression(bquote(italic(d)~"="~.(D_label[[3]])))) # d = 0.1 wlab <- c(as.expression(bquote(italic(W)~"="~.(W_label[[1]]))), # W = 0.0 as.expression(bquote(italic(W)~"="~.(W_label[[2]]))), # W = 0.5 as.expression(bquote(italic(W)~"="~.(W_label[[3]])))) # W = 1.0 klab_short <- c(as.expression(bquote(italic(K)~"="~.(K_label[[1]]))), # K = 0 as.expression(bquote(italic(K)~"="~.(K_label[[2]])))) # K = 1 klab <- c("Scaling: No","Scaling: Yes") labs_short <- expand.grid(K_label, W_label, D_label) # labs <- expand.grid(klab_short, wlab, dlab) # labs_comb <- sprintf('%s; %s; %s', labs[,3], labs[,2], labs[,1])lab_Ar <- "Mean allelic richness" labs_comb <- c(as.expression(bquote(italic(d)~"="~.(D_label[[1]])*";"~italic(W)~"="~.(W_label[[1]])*";"~italic(K)~"="~.(K_label[[1]]))), as.expression(bquote(italic(d)~"="~.(D_label[[1]])*";"~italic(W)~"="~.(W_label[[1]])*";"~italic(K)~"="~.(K_label[[2]]))), as.expression(bquote(italic(d)~"="~.(D_label[[1]])*";"~italic(W)~"="~.(W_label[[2]])*";"~italic(K)~"="~.(K_label[[1]]))), as.expression(bquote(italic(d)~"="~.(D_label[[1]])*";"~italic(W)~"="~.(W_label[[2]])*";"~italic(K)~"="~.(K_label[[2]]))), as.expression(bquote(italic(d)~"="~.(D_label[[1]])*";"~italic(W)~"="~.(W_label[[3]])*";"~italic(K)~"="~.(K_label[[1]]))), as.expression(bquote(italic(d)~"="~.(D_label[[1]])*";"~italic(W)~"="~.(W_label[[3]])*";"~italic(K)~"="~.(K_label[[2]]))), as.expression(bquote(italic(d)~"="~.(D_label[[2]])*";"~italic(W)~"="~.(W_label[[1]])*";"~italic(K)~"="~.(K_label[[1]]))), as.expression(bquote(italic(d)~"="~.(D_label[[2]])*";"~italic(W)~"="~.(W_label[[1]])*";"~italic(K)~"="~.(K_label[[2]]))), as.expression(bquote(italic(d)~"="~.(D_label[[2]])*";"~italic(W)~"="~.(W_label[[2]])*";"~italic(K)~"="~.(K_label[[1]]))), as.expression(bquote(italic(d)~"="~.(D_label[[2]])*";"~italic(W)~"="~.(W_label[[2]])*";"~italic(K)~"="~.(K_label[[2]]))), as.expression(bquote(italic(d)~"="~.(D_label[[2]])*";"~italic(W)~"="~.(W_label[[3]])*";"~italic(K)~"="~.(K_label[[1]]))), as.expression(bquote(italic(d)~"="~.(D_label[[2]])*";"~italic(W)~"="~.(W_label[[3]])*";"~italic(K)~"="~.(K_label[[2]]))), as.expression(bquote(italic(d)~"="~.(D_label[[3]])*";"~italic(W)~"="~.(W_label[[1]])*";"~italic(K)~"="~.(K_label[[1]]))), as.expression(bquote(italic(d)~"="~.(D_label[[3]])*";"~italic(W)~"="~.(W_label[[1]])*";"~italic(K)~"="~.(K_label[[2]]))), as.expression(bquote(italic(d)~"="~.(D_label[[3]])*";"~italic(W)~"="~.(W_label[[2]])*";"~italic(K)~"="~.(K_label[[1]]))), as.expression(bquote(italic(d)~"="~.(D_label[[3]])*";"~italic(W)~"="~.(W_label[[2]])*";"~italic(K)~"="~.(K_label[[2]]))), as.expression(bquote(italic(d)~"="~.(D_label[[3]])*";"~italic(W)~"="~.(W_label[[3]])*";"~italic(K)~"="~.(K_label[[1]]))), as.expression(bquote(italic(d)~"="~.(D_label[[3]])*";"~italic(W)~"="~.(W_label[[3]])*";"~italic(K)~"="~.(K_label[[2]])))) lab_Ar <- "Mean allelic richness" lab_Ho <- "Mean observed heterozygosity" lab_Hs <- "Expected heterozygosity" lab_Ar_short <- "Mean Ar" lab_Ho_short <- "Mean Ho" lab_Hs_short <- "He" measure1 <- "SPO = Sum of perpendicular offsets" measure1_short <- "SPO" measure2 <- "MPO = Median of perpendicular offsets" measure2_short <- "MPO" measure3 <- "Perpendicular offset" measure3a <- "perpendicular offsets" measure4 <- "DMPO" lab_sub <- c("(a)","(b)","(c)","(d)","(e)","(f)") sum_digits <- 1 sum_cex <- 1.4 median_digits <- 3 median_cex <- 1.4 ##*** Colors #### # Colors for two species col_Gfos_A <- "#bf812d" # RGB 191, 129, 45; yellowish or orange col_Gfos_B <- "#35978f" # RGB 53, 151, 143; turquoise # Color for Gammarus fossarum complex col_Gfos <- "#9970ab" # Color for model data colMod <- rgb(70,70,70,max=255) # Colors for waterways col_water <- "darkgrey" col_rhine <- "lightgrey" alpha1 <- "CC" # 80% alpha, check https://gist.github.com/lopspower/03fb1cc0ac9f32ef38f4 for transparency code alpha2 <- "80" # 50% alpha, check https://gist.github.com/lopspower/03fb1cc0ac9f32ef38f4 for transparency code white_transparent <- paste0("#FFFFFF",alpha2) # Transparent white # Preparation for heatmap style color palette col_pal <- c('dark red','white','navy blue') col_fun <- colorRamp(col_pal) rgb2hex <- function(r,g,b) rgb(r, g, b, maxColorValue = 255) paletteLength <- 100 # how finely should the color ramp be divided my_palette <- colorRampPalette(col_pal)(n = paletteLength) #* Functions #### #### Function for GLM and prediction appending to data # The function requires a model formulation, the data, a name for the output, the model family, a significance level, and if selection should be implemented glm.bind <- function(model, data, name, family, sign, step=F){ assign(paste0("glm_",name),glm(formula(model), data, family=family)) b <- get(paste0("glm_",name)) if (step){ assign(paste0("sglm_",name), step(get(paste0("glm_",name)))) # backward selection c <- get(paste0("sglm_",name)) assign(paste0("Predict_sglm_",name),predict(get(paste0("sglm_",name)), type="response", se.fit=T)) assign(paste0("sglm_",name,"_upr"),get(paste0("Predict_sglm_",name))$fit + (sign * get(paste0("Predict_sglm_",name))$se.fit)) assign(paste0("sglm_",name,"_lwr"),get(paste0("Predict_sglm_",name))$fit - (sign * get(paste0("Predict_sglm_",name))$se.fit)) assign(paste0("sglm_",name,"_fit"),get(paste0("Predict_sglm_",name))$fit) data <- cbind(data, cbind(get(paste0("sglm_",name,"_fit")), get(paste0("sglm_",name,"_lwr")), get(paste0("sglm_",name,"_upr")))) colnames(data)[c(ncol(data)-2,ncol(data)-1,ncol(data))] <- c(paste0("sglm_",name,"_fit"),paste0("sglm_",name,"_lwr"),paste0("sglm_",name,"_upr")) }else{ assign(paste0("Predict_glm_",name),predict(get(paste0("glm_",name)), type="response", se.fit=T)) assign(paste0("glm_",name,"_upr"),get(paste0("Predict_glm_",name))$fit + (sign * get(paste0("Predict_glm_",name))$se.fit)) assign(paste0("glm_",name,"_lwr"),get(paste0("Predict_glm_",name))$fit - (sign * get(paste0("Predict_glm_",name))$se.fit)) assign(paste0("glm_",name,"_fit"),get(paste0("Predict_glm_",name))$fit) data <- cbind(data, cbind(get(paste0("glm_",name,"_fit")), get(paste0("glm_",name,"_lwr")), get(paste0("glm_",name,"_upr")))) colnames(data)[c(ncol(data)-2,ncol(data)-1,ncol(data))] <- c(paste0("glm_",name,"_fit"),paste0("glm_",name,"_lwr"),paste0("glm_",name,"_upr")) } a <- data if (step){ return(list(a,b,c)) }else{ return(list(a,b)) } } # The function requires a model formulation, the data, a name for the output, the model family, a significance level, and if selection should be implemented lm.bind <- function(model, data, name, sign, step=F){ assign(paste0("lm_",name),lm(formula(model), data)) b <- get(paste0("lm_",name)) if (step){ assign(paste0("slm_",name), step(get(paste0("lm_",name)))) # backward selection c <- get(paste0("slm_",name)) assign(paste0("Predict_slm_",name),predict(get(paste0("slm_",name)), type="response", se.fit=T)) assign(paste0("slm_",name,"_upr"),get(paste0("Predict_slm_",name))$fit + (sign * get(paste0("Predict_slm_",name))$se.fit)) assign(paste0("slm_",name,"_lwr"),get(paste0("Predict_slm_",name))$fit - (sign * get(paste0("Predict_slm_",name))$se.fit)) assign(paste0("slm_",name,"_fit"),get(paste0("Predict_slm_",name))$fit) data <- cbind(data, cbind(get(paste0("slm_",name,"_fit")), get(paste0("slm_",name,"_lwr")), get(paste0("slm_",name,"_upr")))) colnames(data)[c(ncol(data)-2,ncol(data)-1,ncol(data))] <- c(paste0("slm_",name,"_fit"),paste0("slm_",name,"_lwr"),paste0("slm_",name,"_upr")) }else{ assign(paste0("Predict_lm_",name),predict(get(paste0("lm_",name)), type="response", se.fit=T)) assign(paste0("lm_",name,"_upr"),get(paste0("Predict_lm_",name))$fit + (sign * get(paste0("Predict_lm_",name))$se.fit)) assign(paste0("lm_",name,"_lwr"),get(paste0("Predict_lm_",name))$fit - (sign * get(paste0("Predict_lm_",name))$se.fit)) assign(paste0("lm_",name,"_fit"),get(paste0("Predict_lm_",name))$fit) data <- cbind(data, cbind(get(paste0("lm_",name,"_fit")), get(paste0("lm_",name,"_lwr")), get(paste0("lm_",name,"_upr")))) colnames(data)[c(ncol(data)-2,ncol(data)-1,ncol(data))] <- c(paste0("lm_",name,"_fit"),paste0("lm_",name,"_lwr"),paste0("lm_",name,"_upr")) } a <- data if (step){ return(list(a,b,c)) }else{ return(list(a,b)) } } #### Function for GLM plots # wrapper function to prepare the figures of the GLM output, including confidence intervals (as defined in glm.bind(sign=x)) GLMplot <- function(x,y,dat,model,xlabel,ylabel,ylim=NULL,axislog="", col1=col_Gfos_A, col2=col_Gfos_B, pt.cex=1, CI_border=T, trans=0.3, xrev=F, xax=NULL, yax=NULL, pointtrans=F, cex.lab=2, cex.axis=1.5, legend=T, cex.legend=1.5){ col1trans <- rgb(col2rgb(col1)[1,]/255,col2rgb(col1)[2,]/255,col2rgb(col1)[3,]/255,trans) col2trans <- rgb(col2rgb(col2)[1,]/255,col2rgb(col2)[2,]/255,col2rgb(col2)[3,]/255,trans) form <- reformulate(y, response = x) formfit <- reformulate(y, response = paste0(model,"_fit")) formupr <- reformulate(y, response = paste0(model,"_upr")) formlwr <- reformulate(y, response = paste0(model,"_lwr")) xcol <- which(colnames(dat)==x) ycol <- which(colnames(dat)==y) lwrcol <- which(colnames(dat)==paste0(model,"_lwr")) uprcol <- which(colnames(dat)==paste0(model,"_upr")) DATAordered <- dat[order(dat[,ycol]),] left <- min(DATAordered[,ycol]) right <- max(DATAordered[,ycol]) if (xrev==T){ xrange <- c(right,left) }else{ xrange <- c(left,right) } if (pointtrans==T){ col1point <- col1trans col2point <- col2trans }else{ col1point <- col1 col2point <- col2 } par(mar=c(3.1+cex.lab, 3.1+cex.lab, 0.5, 0.5)) plot(form, dat, type = "n", las = 1, bty = "l", xlab=xlabel, ylab=ylabel, xlim=xrange, ylim=ylim, log=axislog, xaxt=xax, yaxt=yax, cex.lab=cex.lab, cex.axis=cex.axis) polygon(c(rev(DATAordered[,ycol][DATAordered$spec=="A"]), DATAordered[,ycol][DATAordered$spec=="A"]), c(rev(DATAordered[,lwrcol][DATAordered$spec=="A"]), DATAordered[,uprcol][DATAordered$spec=="A"]), col = col1trans, border = NA) polygon(c(rev(DATAordered[,ycol][DATAordered$spec=="B"]), DATAordered[,ycol][DATAordered$spec=="B"]), c(rev(DATAordered[,lwrcol][DATAordered$spec=="B"]), DATAordered[,uprcol][DATAordered$spec=="B"]), col = col2trans, border = NA) points(form, data = subset(DATAordered, spec == "A"), pch = 16, col = col1point, cex=pt.cex) points(form, data = subset(DATAordered, spec == "B"), pch = 16, col = col2point, cex=pt.cex) lines(formfit, data = subset(DATAordered, spec == "A"), lwd = 2.5, col=col1) if(CI_border){lines(formupr, data = subset(DATAordered, spec == "A"), lwd = 2, lty=2, col=col1)} if(CI_border){lines(formlwr, data = subset(DATAordered, spec == "A"), lwd = 2, lty=2, col=col1)} lines(formfit, data = subset(DATAordered, spec == "B"), lwd = 2.5, col=col2) if(CI_border){lines(formupr, data = subset(DATAordered, spec == "B"), lwd = 2, lty=2, col=col2)} if(CI_border){lines(formlwr, data = subset(DATAordered, spec == "B"), lwd = 2, lty=2, col=col2)} if (legend){ legend("bottomright",c(label_A,label_B),pch = 16, col = c(col1,col2), bty="n", cex=cex.legend) } } #### Function to calculate Euclidean distance # As proposed by user Shambho (https://stackoverflow.com/users/3547167) here: https://stackoverflow.com/a/24747155 euc.dist <- function(x1, x2){sqrt(sum((x1 - x2) ^ 2))} #### Function to find endpoint for a perpendicular segment from the point (x0,y0) to the line # As proposed by user MrFlick (https://stackoverflow.com/users/2372064) here: https://stackoverflow.com/a/30399576 perp.segment.coord <- function(x0, y0, a=0,b=1){ # defined by lm.mod as y=a+b*x x1 <- (x0+b*y0-a*b)/(1+b^2) y1 <- a + b*x1 return(list(x0=x0, y0=y0, x1=x1, y1=y1)) } #### Function to calculate decimal variant of ceiling() # As proposed by user Ferroao (https://stackoverflow.com/users/6388753/ferroao) here: https://stackoverflow.com/a/59861612/6380381 ceiling_dec <- function(x, decimals=1) { x2<-x*10^decimals ceiling(x2)/10^decimals } #### Function to specify decimals # Copied from https://stackoverflow.com/a/12135122/6380381 specify_decimal <- function(x, k, formatC=TRUE){ if (formatC){ trimws(formatC(round(x, digits=k),digits=k, format="f")) }else{ trimws(format(round(x, k), nsmall=k)) } } #### DATA PREPARATION -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ #### updist_Mod <- updist[match(modsite,V(net)$name)] catch_Mod <- V(net)$Total_Catch[match(modsite,V(net)$name)] catch_A_red <- V(net)$Total_Catch[match(microsite_A_red,V(net)$name)] # prepare for Gfos A as well (not preexisting as such) catch_B <- V(net)$Total_Catch[match(microsite_B,V(net)$name)] # prepare for Gfos B as well (not preexisting as such) betw_undir_Mod <- betw_undir[match(modsite,V(net)$name)] clos_undir_Mod <- clos_undir[match(modsite,V(net)$name)] degree_undir_Mod <- degree_undir$res[match(modsite,V(net)$name)] if(!existing_data){ ##*** PopGen table preparation #### Ar_modelled <- data.frame(matrix(nrow=length(empiricaldata), ncol=length(D)*length(W)*length(K)+1)) rownames(Ar_modelled) <- adj.mat.names[empiricaldata] Ho_modelled <- data.frame(matrix(nrow=length(empiricaldata), ncol=length(D)*length(W)*length(K)+1)) rownames(Ho_modelled) <- adj.mat.names[empiricaldata] Hs_modelled <- data.frame(matrix(nrow=length(empiricaldata), ncol=length(D)*length(W)*length(K)+1)) rownames(Hs_modelled) <- adj.mat.names[empiricaldata] r <- 0 for (d in 1:length(D)){ # looping over dispersal rates for (w in 1:length(W)){ # looping over dispersal directionalities for (k in 1:length(K)){ # looping over carrying capacities r <- r+1 label_Mod <- paste0("D=",D_label[d],", W_up=",W_label[w],", K=",K_label[k]) load(paste0(DF,"/02_Data_prep/",prep_folder,"/IndPopGenData_",D[d],"_",W[w],"_",K[k],".Rdata")) orderMod <- order(updist_Mod, decreasing = T) if (r==1){ match_Mod <- match_Mod[orderMod] } #*** Mean Ar by upstream distance #### updist_Mod_plot <- updist_Mod[orderMod] meanAr_Mod_updist <- meanAr_Mod[orderMod] orderA_red <- order(updist_A_red, decreasing = T) updist_A_red_plot <- updist_A_red[orderA_red] meanAr_A_red_updist <- meanAr_A_red[orderA_red] orderB <- order(updist_B, decreasing = T) updist_B_plot <- updist_B[orderB] meanAr_B_updist <- meanAr_B[orderB] #*** Mean Ho by upstream distance #### updist_Mod_plot <- updist_Mod[orderMod] meanHo_Mod_updist <- meanHo_Mod[orderMod] orderA_red <- order(updist_A_red, decreasing = T) updist_A_red_plot <- updist_A_red[orderA_red] meanHo_A_red_updist <- meanHo_A_red[orderA_red] orderB <- order(updist_B, decreasing = T) updist_B_plot <- updist_B[orderB] meanHo_B_updist <- meanHo_B[orderB] #*** He by upstream distance #### updist_Mod_plot <- updist_Mod[orderMod] meanHs_Mod_updist <- meanHs_Mod[orderMod] orderA_red <- order(updist_A_red, decreasing = T) updist_A_red_plot <- updist_A_red[orderA_red] meanHs_A_red_updist <- meanHs_A_red[orderA_red] orderB <- order(updist_B, decreasing = T) updist_B_plot <- updist_B[orderB] meanHs_B_updist <- meanHs_B[orderB] # save modelled popgen values to table if (r==1){ Ar_modelled[,1] <- updist_Mod_plot colnames(Ar_modelled)[1] <- "upstream_distance" } Ar_modelled[,1+r] <- meanAr_Mod_updist colnames(Ar_modelled)[r+1] <- paste0(D[d],"_",W[w],"_",K[k]) # save modelled popgen values to table if (r==1){ Ho_modelled[,1] <- updist_Mod_plot colnames(Ho_modelled)[1] <- "upstream_distance" } Ho_modelled[,1+r] <- meanHo_Mod_updist colnames(Ho_modelled)[r+1] <- paste0(D[d],"_",W[w],"_",K[k]) # save modelled popgen values to table if (r==1){ Hs_modelled[,1] <- updist_Mod_plot colnames(Hs_modelled)[1] <- "upstream_distance" } Hs_modelled[,1+r] <- meanHs_Mod_updist colnames(Hs_modelled)[r+1] <- paste0(D[d],"_",W[w],"_",K[k]) } # end looping over carrying capacities } # end looping over dispersal directionalities } # end looping over dispersal rates }else{ load("PopGenData.RData") } #### EXPLANATORY VARIABLES -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ #### # check for correlations expl.var1 <- DATA[,!(colnames(DATA) %in% c("meanAr","meanHo","meanHs","network_match","spec","log_updist","log_catch","clos_undir","clos_dir"))] cor.var1 <- cor(expl.var1, method = "kendall") # transform explanatory variables if(log_trans){ DATA[,c("betw_undir","betw_dir","degree","catch")] <- log(DATA[c("betw_undir","betw_dir","degree","catch")]) DATA[DATA==-Inf] <- 0 } expl.var <- DATA[,!(colnames(DATA) %in% c("meanAr","meanHo","meanHs","degree","betw_undir","clos_dir","network_match","spec","log_updist","log_catch","clos_undir","std_clos_dir"))] cor.var <- cor(expl.var, method = "kendall") #### MODELLING EMPIRICAL DATA -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ #### ##*** Model settings #### #### Define ranges for predict function range_betw_undir_B <- seq(min(betw_undir_B),max(betw_undir_B),(max(betw_undir_B)-min(betw_undir_B))/1000) range_betw_undir_A_red <- seq(min(betw_undir_A_red),max(betw_undir_A_red),(max(betw_undir_A_red)-min(betw_undir_A_red))/1000) range_clos_undir_B <- seq(min(clos_undir_B),max(clos_undir_B),(max(clos_undir_B)-min(clos_undir_B))/1000) range_clos_undir_A_red <- seq(min(clos_undir_A_red),max(clos_undir_A_red),(max(clos_undir_A_red)-min(clos_undir_A_red))/1000) range_degree_undir_B <- seq(min(degree_undir_B),max(degree_undir_B),(max(degree_undir_B)-min(degree_undir_B))/1000) range_degree_undir_A_red <- seq(min(degree_undir_A_red),max(degree_undir_A_red),(max(degree_undir_A_red)-min(degree_undir_A_red))/1000) range_betw_dir_B <- seq(min(betw_dir_B),max(betw_dir_B),(max(betw_dir_B)-min(betw_dir_B))/1000) range_betw_dir_A_red <- seq(min(betw_dir_A_red),max(betw_dir_A_red),(max(betw_dir_A_red)-min(betw_dir_A_red))/1000) range_clos_dir_B <- seq(min(clos_dir_B),max(clos_dir_B),(max(clos_dir_B)-min(clos_dir_B))/1000) range_clos_dir_A_red <- seq(min(clos_dir_A_red),max(clos_dir_A_red),(max(clos_dir_A_red)-min(clos_dir_A_red))/1000) range_updist_B <- seq(1,max(updist_B),1000) range_updist_A_red <- seq(1,max(updist_A_red),1000) range_dist_B <- seq(1,max(dist_B),1000) range_dist_A_red <- seq(1,max(dist_A_red),1000) hist(DATA$catch) ##*** Models meanAr #### #### Initial model exploration shapiro.test(DATA$meanAr) # Interaction model int.mod.Ar <- lm(meanAr ~ updist*betw_dir*std_clos_undir*catch*spec, DATA, na.action = "na.fail") summary(int.mod.Ar) opar <- par(mfrow=c(2,2)) plot(int.mod.Ar) par(opar) step(int.mod.Ar) # best model keeps some interactions if(log_trans){ int.sel.mod.Ar <- lm(meanAr ~ updist + betw_dir + std_clos_undir + catch + spec + updist:betw_dir + updist:std_clos_undir + betw_dir:std_clos_undir + updist:catch + betw_dir:catch + std_clos_undir:catch + updist:spec + betw_dir:spec + std_clos_undir:spec + catch:spec + updist:betw_dir:std_clos_undir + updist:betw_dir:catch + updist:std_clos_undir:catch + betw_dir:std_clos_undir:catch + updist:betw_dir:spec + updist:catch:spec + betw_dir:catch:spec + updist:betw_dir:std_clos_undir:catch + updist:betw_dir:catch:spec, data = DATA, na.action = "na.fail") }else{ int.sel.mod.Ar <- lm(meanAr ~ updist + betw_dir + std_clos_undir + catch + spec + updist:betw_dir + updist:std_clos_undir + betw_dir:std_clos_undir + updist:catch + betw_dir:catch + std_clos_undir:catch + updist:spec + betw_dir:spec + std_clos_undir:spec + catch:spec + updist:betw_dir:std_clos_undir + updist:betw_dir:catch + betw_dir:std_clos_undir:catch + updist:betw_dir:spec + updist:std_clos_undir:spec + betw_dir:std_clos_undir:spec + updist:catch:spec + betw_dir:catch:spec + std_clos_undir:catch:spec + updist:betw_dir:catch:spec + betw_dir:std_clos_undir:catch:spec, data = DATA, na.action = "na.fail") } # Linear model lin.mod.Ar <- lm(meanAr ~ updist+betw_dir+std_clos_undir+catch+spec, DATA, na.action = "na.fail") summary(lin.mod.Ar) opar <- par(mfrow=c(2,2)) plot(lin.mod.Ar) par(opar) MuMIn::dredge(lin.mod.Ar) step(lin.mod.Ar) # best model: ctc+upd (untransformed); btw_dir+spc+upd (log-transformed) if(log_trans){ sel.mod.Ar <- lm(meanAr ~ updist+betw_dir+std_clos_undir, DATA, na.action = "na.fail") }else{ sel.mod.Ar <- lm(meanAr ~ updist+catch, DATA, na.action = "na.fail") } # Comparison AIC(int.sel.mod.Ar) AIC(sel.mod.Ar) AIC(glm(meanAr ~ updist+catch+spec, DATA, family="Gamma")) # linear model outperforms interaction model summary(sel.mod.Ar) opar <- par(mfrow=c(2,2)) plot(sel.mod.Ar) par(opar) car::vif(sel.mod.Ar) #### Full LM of allelic richness without interactions model <- lm.bind(meanAr ~ updist+betw_dir+std_clos_undir+catch+spec, DATA, "Ar_full", critval, step=T) DATA <- model[[1]] lm_Ar_full <- model[[2]] summary(lm_Ar_full) car::vif(lm_Ar_full) slm_Ar_full <- model[[3]] summary(slm_Ar_full) car::vif(slm_Ar_full) # should be the same as above (car::vif(sel.mod.Ar)) #### LM of allelic richness by upstream distance * species AIC(lm(meanAr ~ updist+spec, DATA)) AIC(lm(meanAr ~ updist, DATA)) model <- lm.bind(meanAr ~ updist, DATA, "Ar_updist", critval, step=T) DATA <- model[[1]] lm_Ar_updist <- model[[2]] slm_Ar_updist <- model[[3]] #### LM of allelic richness by undirected closeness centrality AIC(lm(meanAr ~ std_clos_undir+spec, DATA)) AIC(lm(meanAr ~ std_clos_undir, DATA)) model <- lm.bind(meanAr ~ std_clos_undir, DATA, "Ar_clos", critval, step=T) DATA <- model[[1]] lm_Ar_clos <- model[[2]] slm_Ar_clos <- model[[3]] #### LM of allelic richness by directed betweenness centrality AIC(lm(meanAr ~ betw_dir+spec, DATA)) AIC(lm(meanAr ~ betw_dir, DATA)) model <- lm.bind(meanAr ~ betw_dir, DATA, "Ar_betw_dir", critval, step=T) DATA <- model[[1]] lm_Ar_betw_dir <- model[[2]] slm_Ar_betw_dir <- model[[3]] ##*** Models meanHo #### shapiro.test(DATA$meanHo) # Interaction model int.mod.Ho <- lm(meanHo ~ updist*betw_dir*std_clos_undir*catch*spec, DATA, na.action = "na.fail") summary(int.mod.Ho) opar <- par(mfrow=c(2,2)) plot(int.mod.Ho) par(opar) step(int.mod.Ho) # best model keeps some interactions if(log_trans){ int.sel.mod.Ho <- lm(meanHo ~ updist * betw_dir * std_clos_undir * catch * spec, data = DATA, na.action = "na.fail") }else{ int.sel.mod.Ho <- lm(meanHo ~ updist + betw_dir + std_clos_undir + catch + spec + updist:betw_dir + updist:std_clos_undir + betw_dir:std_clos_undir + updist:catch + betw_dir:catch + std_clos_undir:catch + updist:spec + betw_dir:spec + std_clos_undir:spec + catch:spec + updist:betw_dir:std_clos_undir + betw_dir:std_clos_undir:catch + updist:betw_dir:spec + std_clos_undir:catch:spec, data = DATA, na.action = "na.fail") } # Linear model lin.mod.Ho <- lm(meanHo ~ updist+betw_dir+std_clos_undir+catch+spec, DATA, na.action = "na.fail") summary(lin.mod.Ho) opar <- par(mfrow=c(2,2)) plot(lin.mod.Ho) par(opar) MuMIn::dredge(lin.mod.Ho) step(lin.mod.Ho) # best model: btw_dir+spc+std_cls_und (untransformed); btw_dir+spc+std_cls_und (log-transformed) if(log_trans){ sel.mod.Ho <- lm(meanHo ~ betw_dir+std_clos_undir, DATA, na.action = "na.fail") }else{ sel.mod.Ho <- lm(meanHo ~ betw_dir+std_clos_undir+spec, DATA, na.action = "na.fail") } # Comparison AIC(int.sel.mod.Ho) AIC(sel.mod.Ho) # AIC(glm(meanHo ~ betw_dir+std_clos_undir+spec, DATA, family= "quasibinomial")) # linear model outperforms interaction model summary(int.sel.mod.Ho) opar <- par(mfrow=c(2,2)) plot(int.sel.mod.Ho) par(opar) car::vif(int.sel.mod.Ho) summary(sel.mod.Ho) opar <- par(mfrow=c(2,2)) plot(sel.mod.Ho) par(opar) car::vif(sel.mod.Ho) #### Full LM of mean observed heterozygosity without interactions model <- lm.bind(meanHo ~ updist+betw_dir+std_clos_undir+catch+spec, DATA, "Ho_full", critval, step=T) DATA <- model[[1]] lm_Ho_full <- model[[2]] summary(lm_Ho_full) car::vif(lm_Ho_full) slm_Ho_full <- model[[3]] summary(slm_Ho_full) car::vif(slm_Ho_full) # should be the same as above (car::vif(sel.mod.Ho)) #### LM of observed heterozygosity by upstream distance * species AIC(lm(meanHo ~ updist+spec, DATA)) AIC(lm(meanHo ~ updist, DATA)) model <- lm.bind(meanHo ~ updist, DATA, "Ho_updist", critval, step=F) DATA <- model[[1]] lm_Ho_updist <- model[[2]] # slm_Ho_updist <- model[[3]] #### LM of observed heterozygosity by closeness centrality * species AIC(lm(meanHo ~ std_clos_undir+spec, DATA)) AIC(lm(meanHo ~ std_clos_undir, DATA)) model <- lm.bind(meanHo ~ std_clos_undir, DATA, "Ho_clos", critval, step=T) DATA <- model[[1]] lm_Ho_clos <- model[[2]] slm_Ho_clos <- model[[3]] #### LM of observed heterozygosity by directed betweenness centrality * species AIC(lm(meanHo ~ betw_dir+spec, DATA)) AIC(lm(meanHo ~ betw_dir, DATA)) model <- lm.bind(meanHo ~ betw_dir, DATA, "Ho_betw_dir", critval, step=T) DATA <- model[[1]] lm_Ho_betw_dir <- model[[2]] slm_Ho_betw_dir <- model[[3]] ##*** Models He #### shapiro.test(DATA$meanHs) # Interaction model int.mod.Hs <- lm(meanHs ~ updist*betw_dir*std_clos_undir*catch*spec, DATA, na.action = "na.fail") summary(int.mod.Hs) opar <- par(mfrow=c(2,2)) plot(int.mod.Hs) par(opar) step(int.mod.Hs) # best model keeps some interactions if(log_trans){ int.sel.mod.Hs <- lm(meanHs ~ betw_dir + spec + betw_dir:spec, data = DATA, na.action = "na.fail") }else{ int.sel.mod.Hs <- lm(meanHs ~ catch, data = DATA, na.action = "na.fail") } # Linear model lin.mod.Hs <- lm(meanHs ~ updist+betw_dir+std_clos_undir+catch+spec, DATA, na.action = "na.fail") summary(lin.mod.Hs) opar <- par(mfrow=c(2,2)) plot(lin.mod.Hs) par(opar) MuMIn::dredge(lin.mod.Hs) step(lin.mod.Hs) # best model: ctc (untransformed); (Int) (log-transformed); ctc+sps (log-transformed, step) if(log_trans){ sel.mod.Hs <- lm(meanHs ~ 1, DATA, na.action = "na.fail") }else{ sel.mod.Hs <- lm(meanHs ~ catch, DATA, na.action = "na.fail") } # Comparison AIC(int.sel.mod.Hs) AIC(sel.mod.Hs) # AIC(glm(meanHs ~ catch, DATA, family= "quasibinomial")) # linear model outperforms interaction model summary(sel.mod.Hs) opar <- par(mfrow=c(2,2)) plot(sel.mod.Hs) par(opar) car::vif(sel.mod.Hs) #### Full LM of expected heterozygosity without interactions model <- lm.bind(meanHs ~ updist+betw_dir+std_clos_undir+catch+spec, DATA, "Hs_full", critval, step=T) DATA <- model[[1]] lm_Hs_full <- model[[2]] summary(lm_Hs_full) car::vif(lm_Hs_full) slm_Hs_full <- model[[3]] summary(slm_Hs_full) car::vif(slm_Hs_full) # should be the same as above (car::vif(sel.mod.Hs)) #### LM of expected heterozygosity by upstream distance * species AIC(lm(meanHs ~ updist+spec, DATA)) AIC(lm(meanHs ~ updist, DATA)) model <- lm.bind(meanHo ~ updist, DATA, "Hs_updist", critval, step=F) DATA <- model[[1]] lm_Hs_updist <- model[[2]] # slm_Hs_updist <- model[[3]] #### LM of expected heterozygosity by closeness centrality * species AIC(lm(meanHs ~ std_clos_undir+spec, DATA)) AIC(lm(meanHs ~ std_clos_undir, DATA)) model <- lm.bind(meanHs ~ std_clos_undir, DATA, "Hs_clos", critval, step=F) DATA <- model[[1]] lm_Hs_clos <- model[[2]] # slm_Hs_clos <- model[[3]] #### LM of expected heterozygosity by directed betweenness centrality * species AIC(lm(meanHs ~ betw_dir+spec, DATA)) AIC(lm(meanHs ~ betw_dir, DATA)) model <- lm.bind(meanHs ~ betw_dir, DATA, "Hs_betw_dir", critval, step=T) DATA <- model[[1]] lm_Hs_betw_dir <- model[[2]] slm_Hs_betw_dir <- model[[3]] ##*** Models Fst #### #### Spatial distance between populations with Fos A and between populations with Fos B DIST_B <- distances(net, v=V(net)[match(microsite_B,V(net)$name)], to=V(net)[match(microsite_B,V(net)$name)], weights=E(net)) DIST_A_red <- distances(net, v=V(net)[match(microsite_A_red,V(net)$name)], to=V(net)[match(microsite_A_red,V(net)$name)], weights=E(net)) if(fst){ #### Mantel test of genetic differentiation by instream distance mantel_A_red <- vegan::mantel(FST_A_red, DIST_A_red, method="pearson", permutations=1000) mantel_B <- vegan::mantel(FST_B, DIST_B, method="pearson", permutations=1000) #### LM of genetic differentiation by spatial distance model_dist_fst_B <- lm(fst_B~dist_B) summary(model_dist_fst_B) opar <- par(mfrow=c(2,2)) plot(model_dist_fst_B) par(opar) resp_dist_fst_B <- predict(model_dist_fst_B, list(dist_B = range_dist_B), type="response") model_dist_fst_A_red <- lm(fst_A_red~dist_A_red) summary(model_dist_fst_A_red) opar <- par(mfrow=c(2,2)) plot(model_dist_fst_A_red) par(opar) resp_dist_fst_A_red <- predict(model_dist_fst_A_red, list(dist_A_red = range_dist_A_red), type="response") shapiro.test(DISTDATA$nonneg_fst) int.mod.Fst <- lm(nonneg_fst ~ dist*spec, DISTDATA, na.action = "na.fail") summary(int.mod.Fst) opar <- par(mfrow=c(2,2)) plot(int.mod.Fst) par(opar) MuMIn::dredge(int.mod.Fst) step(int.mod.Fst) # interaction model outperforms linear model car::vif(int.mod.Fst) lin.mod.Fst <- lm(nonneg_fst ~ dist+spec, DISTDATA, na.action = "na.fail") summary(lin.mod.Fst) opar <- par(mfrow=c(2,2)) plot(lin.mod.Fst) par(opar) MuMIn::dredge(lin.mod.Fst) car::vif(lin.mod.Fst) log.mod.Fst <- lm(nonneg_fst ~ log(dist)*spec, DISTDATA, na.action = "na.fail") power <- seq(0,1,0.01) AICpower <- c() for (i in 1:length(power)){ pow.mod.Fst <- lm(nonneg_fst ~ I(dist^power[i])*spec, DISTDATA, na.action = "na.fail") AICpower[i] <- AIC(pow.mod.Fst) } plot(AICpower~power, xlab="power term", ylab="AIC") pow.mod.Fst <- lm(nonneg_fst ~ I(dist^power[which.min(AICpower)])*spec, DISTDATA, na.action = "na.fail") AIC(int.mod.Fst) AIC(lin.mod.Fst) AIC(log.mod.Fst) AIC(pow.mod.Fst) summary(pow.mod.Fst) #### Full LM of expected heterozygosity without interactions model <- lm.bind(nonneg_fst ~ I(dist^power[which.min(AICpower)])*spec, DISTDATA, "fst_power", critval, step=T) DISTDATA <- model[[1]] lm_fst_power <- model[[2]] summary(lm_fst_power) car::vif(lm_fst_power) slm_fst_power <- model[[3]] summary(slm_fst_power) car::vif(slm_fst_power) } #### PERPENDICULAR DISTANCES -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ #### ##*** Preparing matrices#### orthdist_Ar_A <- vector() orthdist_Ar_A_directed <- vector() sum_orthdist_Ar_A <- vector() hist_Ar_A <- matrix(nrow=length(modsite_GfosA), ncol=ncol(Ar_modelled)-1) hist_Ar_A_directed <- matrix(nrow=length(modsite_GfosA), ncol=ncol(Ar_modelled)-1) orthdist_Ar_B <- vector() orthdist_Ar_B_directed <- vector() sum_orthdist_Ar_B <- vector() hist_Ar_B <- matrix(nrow=length(modsite_GfosB), ncol=ncol(Ar_modelled)-1) hist_Ar_B_directed <- matrix(nrow=length(modsite_GfosB), ncol=ncol(Ar_modelled)-1) orthdist_Ho_A <- vector() orthdist_Ho_A_directed <- vector() sum_orthdist_Ho_A <- vector() hist_Ho_A <- matrix(nrow=length(modsite_GfosA), ncol=ncol(Ho_modelled)-1) hist_Ho_A_directed <- matrix(nrow=length(modsite_GfosA), ncol=ncol(Ho_modelled)-1) orthdist_Ho_B <- vector() orthdist_Ho_B_directed <- vector() sum_orthdist_Ho_B <- vector() hist_Ho_B <- matrix(nrow=length(modsite_GfosB), ncol=ncol(Ho_modelled)-1) hist_Ho_B_directed <- matrix(nrow=length(modsite_GfosB), ncol=ncol(Ho_modelled)-1) orthdist_Hs_A <- vector() orthdist_Hs_A_directed <- vector() sum_orthdist_Hs_A <- vector() hist_Hs_A <- matrix(nrow=length(modsite_GfosA), ncol=ncol(Hs_modelled)-1) hist_Hs_A_directed <- matrix(nrow=length(modsite_GfosA), ncol=ncol(Hs_modelled)-1) orthdist_Hs_B <- vector() orthdist_Hs_B_directed <- vector() sum_orthdist_Hs_B <- vector() hist_Hs_B <- matrix(nrow=length(modsite_GfosB), ncol=ncol(Hs_modelled)-1) hist_Hs_B_directed <- matrix(nrow=length(modsite_GfosB), ncol=ncol(Hs_modelled)-1) ##*** Calculating distances#### for (i in 2:ncol(Ar_modelled)){ Ar_Mod_A <- Ar_modelled[,i][rownames(Ar_modelled)%in%modsite_GfosA] Ar_Mod_B <- Ar_modelled[,i][rownames(Ar_modelled)%in%modsite_GfosB] Ho_Mod_A <- Ho_modelled[,i][rownames(Ho_modelled)%in%modsite_GfosA] Ho_Mod_B <- Ho_modelled[,i][rownames(Ho_modelled)%in%modsite_GfosB] Hs_Mod_A <- Hs_modelled[,i][rownames(Hs_modelled)%in%modsite_GfosA] Hs_Mod_B <- Hs_modelled[,i][rownames(Hs_modelled)%in%modsite_GfosB] for (j in 1:length(meanAr_A_red_updist)){ point <- cbind(meanAr_A_red_updist,Ar_Mod_A)[j,] seg <- unlist(perp.segment.coord(point[1],point[2])) orthdist_Ar_A[j] <- euc.dist(c(seg[1],seg[2]),c(seg[3],seg[4])) orthdist_Ar_A_directed[j] <- sign(seg[2]-seg[4])*orthdist_Ar_A[j] } for (k in 1:length(meanAr_B_updist)){ point <- cbind(meanAr_B_updist,Ar_Mod_B)[k,] seg <- unlist(perp.segment.coord(point[1],point[2])) orthdist_Ar_B[k] <- euc.dist(c(seg[1],seg[2]),c(seg[3],seg[4])) orthdist_Ar_B_directed[k] <- sign(seg[2]-seg[4])*orthdist_Ar_B[k] } for (j in 1:length(meanHo_A_red_updist)){ point <- cbind(meanHo_A_red_updist,Ho_Mod_A)[j,] seg <- unlist(perp.segment.coord(point[1],point[2])) orthdist_Ho_A[j] <- euc.dist(c(seg[1],seg[2]),c(seg[3],seg[4])) orthdist_Ho_A_directed[j] <- sign(seg[2]-seg[4])*orthdist_Ho_A[j] } for (k in 1:length(meanHo_B_updist)){ point <- cbind(meanHo_B_updist,Ho_Mod_B)[k,] seg <- unlist(perp.segment.coord(point[1],point[2])) orthdist_Ho_B[k] <- euc.dist(c(seg[1],seg[2]),c(seg[3],seg[4])) orthdist_Ho_B_directed[k] <- sign(seg[2]-seg[4])*orthdist_Ho_B[k] } for (j in 1:length(meanHs_A_red_updist)){ point <- cbind(meanHs_A_red_updist,Hs_Mod_A)[j,] seg <- unlist(perp.segment.coord(point[1],point[2])) orthdist_Hs_A[j] <- euc.dist(c(seg[1],seg[2]),c(seg[3],seg[4])) orthdist_Hs_A_directed[j] <- sign(seg[2]-seg[4])*orthdist_Ho_A[j] } for (k in 1:length(meanHs_B_updist)){ point <- cbind(meanHs_B_updist,Hs_Mod_B)[k,] seg <- unlist(perp.segment.coord(point[1],point[2])) orthdist_Hs_B[k] <- euc.dist(c(seg[1],seg[2]),c(seg[3],seg[4])) orthdist_Hs_B_directed[k] <- sign(seg[2]-seg[4])*orthdist_Ho_B[k] } sum_orthdist_Ar_A[[i-1]] <- sum(orthdist_Ar_A) sum_orthdist_Ar_B[[i-1]] <- sum(orthdist_Ar_B) sum_orthdist_Ho_A[[i-1]] <- sum(orthdist_Ho_A) sum_orthdist_Ho_B[[i-1]] <- sum(orthdist_Ho_B) sum_orthdist_Hs_A[[i-1]] <- sum(orthdist_Hs_A) sum_orthdist_Hs_B[[i-1]] <- sum(orthdist_Hs_B) hist_Ar_A[,i-1] <- orthdist_Ar_A hist_Ar_B[,i-1] <- orthdist_Ar_B hist_Ho_A[,i-1] <- orthdist_Ho_A hist_Ho_B[,i-1] <- orthdist_Ho_B hist_Hs_A[,i-1] <- orthdist_Hs_A hist_Hs_B[,i-1] <- orthdist_Hs_B hist_Ar_A_directed[,i-1] <- orthdist_Ar_A_directed hist_Ar_B_directed[,i-1] <- orthdist_Ar_B_directed hist_Ho_A_directed[,i-1] <- orthdist_Ho_A_directed hist_Ho_B_directed[,i-1] <- orthdist_Ho_B_directed hist_Hs_A_directed[,i-1] <- orthdist_Hs_A_directed hist_Hs_B_directed[,i-1] <- orthdist_Hs_B_directed } names(sum_orthdist_Ar_A) <- colnames(Ar_modelled)[-1] names(sum_orthdist_Ar_B) <- colnames(Ar_modelled)[-1] names(sum_orthdist_Ho_A) <- colnames(Ho_modelled)[-1] names(sum_orthdist_Ho_B) <- colnames(Ho_modelled)[-1] names(sum_orthdist_Hs_A) <- colnames(Hs_modelled)[-1] names(sum_orthdist_Hs_B) <- colnames(Hs_modelled)[-1] #### VALUES -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ #### nrow(A_red)+nrow(B) # Number of used individuals if (internal){ nrow(Gfos[which(Gfos$type!="K"),]) } min(meanAr_A_red) max(meanAr_A_red) mean(meanAr_A_red) median(meanAr_A_red) sd(meanAr_A_red) min(meanAr_B) max(meanAr_B) mean(meanAr_B) median(meanAr_B) sd(meanAr_B) min(meanHo_A_red) max(meanHo_A_red) mean(meanHo_A_red) median(meanHo_A_red) sd(meanHo_A_red) min(meanHo_B) max(meanHo_B) mean(meanHo_B) median(meanHo_B) sd(meanHo_B) min(meanHs_A_red) max(meanHs_A_red) mean(meanHs_A_red) median(meanHs_A_red) sd(meanHs_A_red) min(meanHs_B) max(meanHs_B) mean(meanHs_B) median(meanHs_B) sd(meanHs_B) if(fst){ print(min(fst_A_red)) print(max(fst_A_red)) print(mean(fst_A_red)) print(median(fst_A_red)) print(sd(fst_A_red)) print(min(fst_B)) print(max(fst_B)) print(mean(fst_B)) print(median(fst_B)) print(sd(fst_B)) print(mantel_A_red) print(mantel_B) } # Get range of scaled carrying capacities (from C++ file, lines 166, 215) # Extract total catchments sizes from graph object (since included here, otherwise use catch_area.in directly) total_catch <- V(net)$Total_Catch # K scaling according to sqrt(catchment size) sqrt(total_catch) sum_sqrt_catch_size <- sum(sqrt(total_catch)) K_scaled <- round(sqrt(total_catch) * (1000*length(total_catch)/sum_sqrt_catch_size )) range(K_scaled) #### FIGURES -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ #### dir.create(paste0(WD,"/Analysis_",output), showWarnings=F) ##** Preparing multipanel figures #### pla <- list( a=2, # columns of subplots b=1, # rows of subplots x=3, # number of rows y=3, # number of cols sub1=F, sub2=T, main=F, main_t=NULL, sub1_t=NULL, sub2_t=klab, x_t=wlab, y_t=dlab, main_c=NULL, sub1_c=3, sub2_c=2.5, x_c=2, y_c=2) lab <- c(pla$sub1_t,rep(c(pla$sub2_t),ifelse(pla$sub2,pla$b,0)),rep(pla$x_t,pla$b),rep(pla$y_t,pla$a)) lab_cex <- c(rep(pla$sub1_c,length(pla$sub1_t)),rep(c(pla$sub2_c),ifelse(pla$sub2,pla$a*pla$b,0)),rep(pla$x_c,pla$b*length(pla$x_t)),rep(pla$y_c,pla$a*length(pla$y_t))) ##** FIG 1 #### ## Empirical data maps if(pdf){ pdf(paste0(WD,"/Analysis_",output,"/Fig1.pdf"), width=1.5*fig.width, height=(3*fig.width)/4.5) }else{ png(paste0(WD,"/Analysis_",output,"/Fig1.png"), width=1.5*fig.width, height=(3*fig.width)/4.5, units="in", res=300) } nf <- layout(matrix(c(12, 7, 8, 9, 10, 1, 2, 3, 11, 4, 5, 6), nrow=3, byrow=T), widths=c(0.5,3,3,3), heights=c(0.5,2,2), respect=T) # layout.show(nf) figmar <- c(4.5,4.5,0.5,0) figmgp <- c(1.7,0.3,0) col_switch <- 1/2 # #### Map of mean allelic richness in G. fossarum A par(mar=figmar, mgp=figmgp, tcl=0.2, xaxs="i", yaxs="i") if (internal){ river_plot(north_arrow = F, overview_map = F, scalebar=T, arrows = F, border_outline=F, width_border=2, col_rhine = col_rhine, col_water=col_water, plot_rhone=F, lines_rhone=F, plot_ticino=F, lines_ticino=F, plot_inn=F, lines_inn=F, lakes=TRUE, rivers=TRUE, axes="no_label", river_nr=FALSE) }else{ river_plot(north_arrow = F, overview_map = F, scalebar=T, arrows = F, border_outline=F, width_border=2, col_rhine = col_rhine, col_water=col_water, plot_rhone=F, plot_ticino=F, plot_inn=F, lakes=TRUE, rivers=TRUE, axes="no_label", river_nr=FALSE) } values <- (meanAr_A_red-min(meanAr_A_red))/ (max(meanAr_A_red)-min(meanAr_A_red)) # transform meanAr values to range [0,1] for heatmap plotting if(!any(is.na(values))){ for(i in 1:length(match_A_red)){ points(site_coord$x[match_A_red[i]], site_coord$y[match_A_red[i]], col=rgb2hex(col_fun(values))[i], pch=19, cex=2.5) l1 <- mean(meanAr_A_red)+(col_switch)*(max(meanAr_A_red)-median(meanAr_A_red)) l2 <- mean(meanAr_A_red)-(col_switch)*(median(meanAr_A_red)-min(meanAr_A_red)) text(site_coord$x[match_A_red[i]], site_coord$y[match_A_red[i]], round(meanAr_A_red[i],1), col=ifelse(meanAr_A_red[i]>l1|meanAr_A_red[i]<l2,"white","black"), cex=0.8) } } overview_map(xl = 495000,xr = 545000,yt = 280000, yb = 230000) north_arrow(x=825000, y=80000) gradient.legend(meanAr_A_red, val.cex = 1, palette=col_pal) mtext(side = 3, text = lab_sub[1], line = 0.5, adj=0, cex = 1.5) #### Map of mean observed heterozygosity in G. fossarum A par(mar=figmar, mgp=figmgp, tcl=0.2, xaxs="i", yaxs="i") if (internal){ river_plot(north_arrow = F, overview_map = F, scalebar=F, arrows = F, border_outline=F, width_border=2, col_rhine = col_rhine, col_water=col_water, plot_rhone=F, lines_rhone=F, plot_ticino=F, lines_ticino=F, plot_inn=F, lines_inn=F, lakes=TRUE, rivers=TRUE, axes="no_label", river_nr=FALSE) }else{ river_plot(north_arrow = F, overview_map = F, scalebar=F, arrows = F, border_outline=F, width_border=2, col_rhine = col_rhine, col_water=col_water, plot_rhone=F, plot_ticino=F, plot_inn=F, lakes=TRUE, rivers=TRUE, axes="no_label", river_nr=FALSE) } mtext(side = 2, text = expression(paste(bold("CH1903"), " / ", bold("LV03"))), line = 1 + 1, cex = 1) values <- (meanHo_A_red-min(meanHo_A_red))/ (max(meanHo_A_red)-min(meanHo_A_red)) # transform meanHo values to range [0,1] for heatmap plotting if(!any(is.na(values))){ for(i in 1:length(match_A_red)){ points(site_coord$x[match_A_red[i]], site_coord$y[match_A_red[i]], col=rgb2hex(col_fun(values))[i], pch=19, cex=2.5) l1 <- mean(meanHo_A_red)+(col_switch)*(max(meanHo_A_red)-median(meanHo_A_red)) l2 <- mean(meanHo_A_red)-(col_switch)*(median(meanHo_A_red)-min(meanHo_A_red)) text(site_coord$x[match_A_red[i]], site_coord$y[match_A_red[i]], round(meanHo_A_red[i],1), col=ifelse(meanHo_A_red[i]>l1|meanHo_A_red[i]<l2,"white","black"), cex=0.8) } } gradient.legend(meanHo_A_red, val.cex=1, palette=col_pal) mtext(side = 3, text = lab_sub[3], line = 0.5, adj=0, cex = 1.5) #### Map of expected heterozygosity in G. fossarum A par(mar=figmar, mgp=figmgp, tcl=0.2, xaxs="i", yaxs="i") if (internal){ river_plot(north_arrow = F, overview_map = F, scalebar=F, arrows = F, border_outline=F, width_border=2, col_rhine = col_rhine, col_water=col_water, plot_rhone=F, lines_rhone=F, plot_ticino=F, lines_ticino=F, plot_inn=F, lines_inn=F, lakes=TRUE, rivers=TRUE, axes="no_label", river_nr=FALSE) }else{ river_plot(north_arrow = F, overview_map = F, scalebar=F, arrows = F, border_outline=F, width_border=2, col_rhine = col_rhine, col_water=col_water, plot_rhone=F, plot_ticino=F, plot_inn=F, lakes=TRUE, rivers=TRUE, axes="no_label", river_nr=FALSE) } mtext(side = 2, text = expression(paste(bold("CH1903"), " / ", bold("LV03"))), line = 1 + 1, cex = 1) values <- (meanHs_A_red-min(meanHs_A_red))/ (max(meanHs_A_red)-min(meanHs_A_red)) # transform meanHs values to range [0,1] for heatmap plotting if(!any(is.na(values))){ for(i in 1:length(match_A_red)){ points(site_coord$x[match_A_red[i]], site_coord$y[match_A_red[i]], col=rgb2hex(col_fun(values))[i], pch=19, cex=2.5) l1 <- mean(meanHs_A_red)+(col_switch)*(max(meanHs_A_red)-median(meanHs_A_red)) l2 <- mean(meanHs_A_red)-(col_switch)*(median(meanHs_A_red)-min(meanHs_A_red)) text(site_coord$x[match_A_red[i]], site_coord$y[match_A_red[i]], round(meanHs_A_red[i],1), col=ifelse(meanHs_A_red[i]>l1|meanHs_A_red[i]<l2,"white","black"), cex=0.8) } } gradient.legend(meanHs_A_red, val.cex=1, palette=col_pal) mtext(side = 3, text = lab_sub[5], line = 0.5, adj=0, cex = 1.5) #### Map of mean allelic richness in G. fossarum B par(mar=figmar, mgp=figmgp, tcl=0.2, xaxs="i", yaxs="i") if (internal){ river_plot(north_arrow = F, overview_map = F, scalebar=F, arrows = F, border_outline=F, width_border=2, col_rhine = col_rhine, col_water=col_water, plot_rhone=F, lines_rhone=F, plot_ticino=F, lines_ticino=F, plot_inn=F, lines_inn=F, lakes=TRUE, rivers=TRUE, axes="no_label", river_nr=FALSE) }else{ river_plot(north_arrow = F, overview_map = F, scalebar=F, arrows = F, border_outline=F, width_border=2, col_rhine = col_rhine, col_water=col_water, plot_rhone=F, plot_ticino=F, plot_inn=F, lakes=TRUE, rivers=TRUE, axes="no_label", river_nr=FALSE) } mtext(side = 1, text = expression(paste(bold("CH1903"), " / ", bold("LV03"))), line = 1 + 1, cex = 1) values <- (meanAr_B-min(meanAr_B))/ (max(meanAr_B)-min(meanAr_B)) # transform meanAr values to range [0,1] for heatmap plotting if(!any(is.na(values))){ for(i in 1:length(match_B)){ points(site_coord$x[match_B[i]], site_coord$y[match_B[i]], col=rgb2hex(col_fun(values))[i], pch=19, cex=2.5) l1 <- mean(meanAr_B)+(col_switch)*(max(meanAr_B)-median(meanAr_B)) l2 <- mean(meanAr_B)-(col_switch)*(median(meanAr_B)-min(meanAr_B)) text(site_coord$x[match_B[i]], site_coord$y[match_B[i]], round(meanAr_B[i],1), col=ifelse(meanAr_B[i]>l1|meanAr_B[i]<l2,"white","black"), cex=0.8) } } gradient.legend(meanAr_B, val.cex = 1, palette=col_pal) mtext(side = 3, text = lab_sub[2], line = 0.5, adj=0, cex = 1.5) #### Map of mean observed heterozygosity in G. fossarum B par(mar=figmar, mgp=figmgp, tcl=0.2, xaxs="i", yaxs="i") if (internal){ river_plot(north_arrow = F, overview_map = F, scalebar=F, arrows = F, border_outline=F, width_border=2, col_rhine = col_rhine, col_water=col_water, plot_rhone=F, lines_rhone=F, plot_ticino=F, lines_ticino=F, plot_inn=F, lines_inn=F, lakes=TRUE, rivers=TRUE, axes="no_label", river_nr=FALSE) }else{ river_plot(north_arrow = F, overview_map = F, scalebar=F, arrows = F, border_outline=F, width_border=2, col_rhine = col_rhine, col_water=col_water, plot_rhone=F, plot_ticino=F, plot_inn=F, lakes=TRUE, rivers=TRUE, axes="no_label", river_nr=FALSE) } mtext(side = 1, text = expression(paste(bold("CH1903"), " / ", bold("LV03"))), line = 1 + 1, cex = 1) mtext(side = 2, text = expression(paste(bold("CH1903"), " / ", bold("LV03"))), line = 1 + 1, cex = 1) values <- (meanHo_B-min(meanHo_B))/ (max(meanHo_B)-min(meanHo_B)) # transform meanHo values to range [0,1] for heatmap plotting for(i in 1:length(match_B)){ points(site_coord$x[match_B[i]], site_coord$y[match_B[i]], col=rgb2hex(col_fun(values))[i], pch=19, cex=2.5) l1 <- mean(meanHo_B)+(col_switch)*(max(meanHo_B)-median(meanHo_B)) l2 <- mean(meanHo_B)-(col_switch)*(median(meanHo_B)-min(meanHo_B)) text(site_coord$x[match_B[i]], site_coord$y[match_B[i]], round(meanHo_B[i],1), col=ifelse(meanHo_B[i]>l1|meanHo_B[i]<l2,"white","black"), cex=0.8) } gradient.legend(meanHo_B, val.cex = 1, palette=col_pal) mtext(side = 3, text = lab_sub[4], line = 0.5, adj=0, cex = 1.5) #### Map of expected heterozygosity in G. fossarum B par(mar=figmar, mgp=figmgp, tcl=0.2, xaxs="i", yaxs="i") if (internal){ river_plot(north_arrow = F, overview_map = F, scalebar=F, arrows = F, border_outline=F, width_border=2, col_rhine = col_rhine, col_water=col_water, plot_rhone=F, lines_rhone=F, plot_ticino=F, lines_ticino=F, plot_inn=F, lines_inn=F, lakes=TRUE, rivers=TRUE, axes="no_label", river_nr=FALSE) }else{ river_plot(north_arrow = F, overview_map = F, scalebar=F, arrows = F, border_outline=F, width_border=2, col_rhine = col_rhine, col_water=col_water, plot_rhone=F, plot_ticino=F, plot_inn=F, lakes=TRUE, rivers=TRUE, axes="no_label", river_nr=FALSE) } mtext(side = 1, text = expression(paste(bold("CH1903"), " / ", bold("LV03"))), line = 1 + 1, cex = 1) mtext(side = 2, text = expression(paste(bold("CH1903"), " / ", bold("LV03"))), line = 1 + 1, cex = 1) values <- (meanHs_B-min(meanHs_B))/ (max(meanHs_B)-min(meanHs_B)) # transform meanHs values to range [0,1] for heatmap plotting for(i in 1:length(match_B)){ points(site_coord$x[match_B[i]], site_coord$y[match_B[i]], col=rgb2hex(col_fun(values))[i], pch=19, cex=2.5) l1 <- mean(meanHs_B)+(col_switch)*(max(meanHs_B)-median(meanHs_B)) l2 <- mean(meanHs_B)-(col_switch)*(median(meanHs_B)-min(meanHs_B)) text(site_coord$x[match_B[i]], site_coord$y[match_B[i]], round(meanHs_B[i],1), col=ifelse(meanHs_B[i]>l1|meanHs_B[i]<l2,"white","black"), cex=0.8) } gradient.legend(meanHs_B, val.cex = 1, palette=col_pal) mtext(side = 3, text = lab_sub[6], line = 0.5, adj=0, cex = 1.5) par(mar=c(0,0,0,0),mgp=c(3,1,0)) plot.new() text(0.5,0.5, lab_Ar ,adj=c(0.5,0.5), cex=2.5) plot.new() text(0.5,0.5, lab_Ho ,adj=c(0.5,0.5), cex=2.5) plot.new() text(0.5,0.5, lab_Hs ,adj=c(0.5,0.5), cex=2.5) plot.new() text(1,0.5,label_A,adj=c(0.5,0), cex=2.5, srt=90) plot.new() text(1,0.5,label_B,adj=c(0.5,0), cex=2.5, srt=90) dev.off() ##** FIG 2 #### if(pdf){ pdf(paste0(WD,"/Analysis_",output,"/Fig2.pdf"), width=22.5, height=10) }else{ png(paste0(WD,"/Analysis_",output,"/Fig2.png"), width=22.5, height=10, units="in", res=300) } par(mfrow=c(2,3)) ylim <- c(min(meanHo_A_red,meanHo_B,meanHs_A_red,meanHs_B),max(meanHo_A_red,meanHo_B,meanHs_A_red,meanHs_B)) #### meanAr~updist GLMplot("meanAr","updist",DATA,model="slm_Ar_updist",pt.cex=2, CI_border=F,xlabel="Upstream distance [km]",ylabel=lab_Ar, xrev=T, xax="n", cex.lab=2, cex.axis=1.5, cex.legend=1.5) axis(1, c(0,50,100,150,200,250,300), at=c(0,50000,100000,150000,200000,250000,300000), cex.axis=1.5) text(300000, par("usr")[4], lab_sub[1], cex=2, adj=c(0.5,1)) #### meanHo~updist GLMplot("meanHo","updist",DATA,model="lm_Ho_updist",ylim=ylim,pt.cex=2, CI_border=F,xlabel="Upstream distance [km]",ylabel=lab_Ho, xrev=T, xax="n", cex.lab=2, cex.axis=1.5, legend=F) axis(1, c(0,50,100,150,200,250,300), at=c(0,50000,100000,150000,200000,250000,300000), cex.axis=1.5) text(300000, par("usr")[4], lab_sub[3], cex=2, adj=c(0.5,1)) #### meanHs~updist GLMplot("meanHs","updist",DATA,model="lm_Hs_updist",ylim=ylim,pt.cex=2, CI_border=F,xlabel="Upstream distance [km]",ylabel=lab_Hs, xrev=T, xax="n", cex.lab=2, cex.axis=1.5, legend=F) axis(1, c(0,50,100,150,200,250,300), at=c(0,50000,100000,150000,200000,250000,300000), cex.axis=1.5) text(300000, par("usr")[4], lab_sub[5], cex=2, adj=c(0.5,1)) #### meanAr~std_clos_undir GLMplot("meanAr","std_clos_undir",DATA,model="slm_Ar_clos",pt.cex=2, CI_border=F,xlabel="Standardized closeness centrality",ylabel=lab_Ar, legend=F) text(0, par("usr")[4], lab_sub[2], cex=2, adj=c(0.5,1)) #### meanHo~std_clos_undir GLMplot("meanHo","std_clos_undir",DATA,model="slm_Ho_clos",ylim=ylim,pt.cex=2, CI_border=F,xlabel="Standardized closeness centrality",ylabel=lab_Ho, legend=F) text(0, par("usr")[4], lab_sub[4], cex=2, adj=c(0.5,1)) # dev.off() #### meanHs~std_clos_undir GLMplot("meanHs","std_clos_undir",DATA,model="lm_Hs_clos",ylim=ylim,pt.cex=2, CI_border=F,xlabel="Standardized closeness centrality",ylabel=lab_Hs, legend=F) text(0, par("usr")[4], lab_sub[6], cex=2, adj=c(0.5,1)) dev.off() ##** FIG 3 #### ##** Modelled data maps #*** Mean Ar maps main_3=F mp_dim <- multipanel.dimensions(main.col=pla$a,main.row=pla$b,sub.row=pla$x,sub.col=pla$y,sub1=pla$sub1,sub2=pla$sub2, main=main_3, h.main=0.5, w.legend=0, h.sub2=0.3, w.axis=0.7, h.axis=0, spacer.sub.col=0.5) if(pdf){ pdf(paste0(WD,"/Analysis_",output,"/Fig3.pdf"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1])) }else{ png(paste0(WD,"/Analysis_",output,"/Fig3.png"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1]), units="in", res=600) } par(mar=c(0,0,0,0)) nf <- multipanel.layout(main.col=pla$a,main.row=pla$b,sub.row=pla$x,sub.col=pla$y,sub1=pla$sub1,sub2=pla$sub2, main=main_3, h.main=0.5, w.legend=0, h.sub2=0.3, w.axis=0.7, h.axis=0, spacer.sub.col=0.5) for (j in 2:ncol(Ar_modelled)){ if (internal){ river_plot(width_country=0.5, lwd_rivers=0.5, lwd_lakes=0.5, xlimit=c(495000,825000), col_water=col_water, north_arrow = F, overview_map = F, scalebar=F, arrows = F, border_outline=F, width_border=2, col_rhine = col_rhine, plot_rhone=F, lines_rhone=F, plot_ticino=F, lines_ticino=F, plot_inn=F, lines_inn=F, lakes=TRUE, rivers=TRUE, axes="none", river_nr=T) }else{ river_plot(width_country=0.5, lwd_rivers=0.5, lwd_lakes=0.5, xlimit=c(495000,825000), col_water=col_water, north_arrow = F, overview_map = F, scalebar=F, arrows = F, border_outline=F, width_border=2, col_rhine = col_rhine, plot_rhone=F, plot_ticino=F, plot_inn=F, lakes=TRUE, rivers=TRUE, axes="none", river_nr=F) } values <- (Ar_modelled[,j]-min(Ar_modelled[,j]))/ (max(Ar_modelled[,j])-min(Ar_modelled[,j])) # transform meanAr values to range [0,1] for heatmap plotting if(!any(is.na(values))){ for(i in 1:length(match_Mod)){ points(site_coord$x[match_Mod[i]], site_coord$y[match_Mod[i]], bg=rgb2hex(col_fun(values))[i], pch=21, lwd=0.5, cex=1) } } if (j %in% c(2,3)){ mtext(lab[6], side=3, line=-1, cex=1.3) } if (j %in% c(8,9)){ mtext(lab[7], side=3, line=-1, cex=1.3) } if (j %in% c(14,15)){ mtext(lab[8], side=3, line=-1, cex=1.3) } gradient.legend(Ar_modelled[,j],alpha=1, val.midpoint=F, round=1, val.cex=1, val.gap = 0.5, title.gap=0.1, xl = 505000, xr = 635000, yb = 20000, yt = 40000, horizontal=T) } if(main_3){ plot.new() text(0.5,0.5,"Simulated mean allelic richness",adj=c(0.5,0.5), cex=3) } for (i in 1:5){ plot.new() if (i %in% c(1:2)){ text(0.5,1,lab[i],adj=c(0.5,1), cex=lab_cex[i]) } if (i %in% c(3:5)){ text(0.5,0.5,lab[i],adj=c(0.5,0.5), cex=lab_cex[i]) } } dev.off() ##** FIG 4 #### ##** Histogram of orthogonal distance to 1:1 line main_4=F mp_dim <- multipanel.dimensions(main.col=pla$a,main.row=pla$b,pla$x,pla$y,sub1=pla$sub1,sub2=pla$sub2,main=main_4, h.main=0.5) if(pdf){ pdf(paste0(WD,"/Analysis_",output,"/Fig4.pdf"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1])) }else{ png(paste0(WD,"/Analysis_",output,"/Fig4.png"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1]), units="in", res=600) } nf <- multipanel.layout(main.col=pla$a,main.row=pla$b,pla$x,pla$y,sub1=pla$sub1,sub2=pla$sub2,main=main_4, h.main=0.5) for (i in 1:ncol(hist_Ar_A)){ par(mar=c(0,0,0,0)) yclip <- 30 ylim <- 40 hist(hist_Ar_A[,i], breaks=seq(0,ceiling(max(hist_Ar_A,hist_Ar_B, na.rm=T))-0.5,0.5), xlim=c(0,ceiling(max(hist_Ar_A,hist_Ar_B, na.rm=T))), ylim=c(0,ylim), xaxt="n", yaxt="n", col=col_Gfos_A, main="") clip(0,ceiling(max(hist_Ar_A,hist_Ar_B, na.rm=T)), 0, yclip) abline(v=median(hist_Ar_A[,i], na.rm=T),col=col_Gfos_A) abline(v=median(hist_Ar_B[,i], na.rm=T),col=col_Gfos_B) clip(0,ceiling(max(hist_Ar_A,hist_Ar_B, na.rm=T)), 0, ylim) textbox(ceiling(max(hist_Ar_A,hist_Ar_B, na.rm=T)),30,paste0(measure2_short," = ",formatC(round(median(hist_Ar_A[,i], na.rm=T),median_digits),digits=median_digits, format="f")), txt.cex=median_cex, txt.col=col_Gfos_A, txt.adj=1, frm.col="white", frm.brd = "white", frm.siz = 0.2) textbox(ceiling(max(hist_Ar_A,hist_Ar_B, na.rm=T)),25,paste0(measure2_short," = ",formatC(round(median(hist_Ar_B[,i], na.rm=T),median_digits),digits=median_digits, format="f")), txt.cex=median_cex, txt.col=col_Gfos_B, txt.adj=1, frm.col="white", frm.brd = "white", frm.siz = 0.2) hist(hist_Ar_A[,i], breaks=seq(0,ceiling(max(hist_Ar_A,hist_Ar_B, na.rm=T))-0.5,0.5), xlim=c(0,ceiling(max(hist_Ar_A,hist_Ar_B, na.rm=T))), ylim=c(0,ylim), xaxt="n", yaxt="n", col=col_Gfos_A, main="", add=T) hist(hist_Ar_B[,i], breaks=seq(0,ceiling(max(hist_Ar_A,hist_Ar_B, na.rm=T))-0.5,0.5), xlim=c(0,ceiling(max(hist_Ar_A,hist_Ar_B, na.rm=T))), ylim=c(0,ylim), xaxt="n", yaxt="n", col=col_Gfos_B, main="", add=T) if (i %in% c(1,2)){ mtext(lab[6], side=3, line=0.8, cex=1.3) } if (i %in% c(7,8)){ mtext(lab[7], side=3, line=0.8, cex=1.3) } if (i %in% c(13,14)){ mtext(lab[8], side=3, line=0.8, cex=1.3) } if (i %in% c(3)){ mtext("Frequencies [counts]", side=2, line=1, cex=1) } if (i %in% c(11,12)){ mtext(measure3, side=1, line=3, cex=1) } if (i %in% c(5:6,11:12,17:18)){ axis(1) } if (i %in% c(1,3,5)){ axis(2, at=seq(0,yclip,5), labels=F) } if (i %in% c(2,4,6)){ axis(2, at=seq(0,yclip,5)) } if (i %in% c(13:18)){ axis(4, at=seq(0,yclip,5), labels=F) } } if(main_4){ plot.new() text(0.5,0.7,paste0(lab_Ar,": Distribution of ",measure3a),adj=c(0.5,0.5),cex=3) } for (i in 1:length(lab)){ plot.new() if (i %in% c(1:2)){ text(0.5,1,lab[i],adj=c(0.5,1), cex=lab_cex[i]) } if (i %in% c(3:5)){ text(0,0.5,lab[i],adj=c(0,0.5), cex=lab_cex[i]) } } # Add example plot par(mar=c(0,6,16,0.5), pty="s") i <- ncol(Ar_modelled) Ar_Mod_A <- Ar_modelled[,i][rownames(Ar_modelled)%in%modsite_GfosA] Ar_Mod_B <- Ar_modelled[,i][rownames(Ar_modelled)%in%modsite_GfosB] plot(Ar_Mod_A~meanAr_A_red_updist, type="n", xlim=c(min(Ar_modelled[,-1], na.rm=T),max(Ar_modelled[,-1], na.rm=T)), ylim=c(min(Ar_modelled[,-1], na.rm=T),max(Ar_modelled[,-1], na.rm=T)), xaxt="s", yaxt="n", xlab="", ylab="", asp=1) par(xpd=T) polygon(c(-4.5,-4.5,par("usr")[3],par("usr")[3]),c(par("usr")[1],9,par("usr")[4],par("usr")[1]), col="lightgrey", border=NA) segments(-4.5,par("usr")[1],par("usr")[3],par("usr")[1], lty=1, col="grey", lwd=1.5) segments(-4.5,9,par("usr")[3],par("usr")[4], lty=1, col="grey", lwd=1.5) text(-2.5,18,"Example plot for\nperpendicular offsets", cex=2, adj=0, col="darkgrey") text(-2.5,15,"See Fig. S10 for all plots",cex=1, adj=0, col="darkgrey") par(xpd=F) mtext(expression(bold("Model data:")*" Ar"), side=2, line=2, cex=0.8) mtext(expression(bold("Empirical data:")*" Ar"), side=1, line=2.5, cex=0.8) box() axis(3,labels=F) axis(2) abline(0,1, lwd=1, lty=2) # add 1:1 line points(Ar_Mod_A~meanAr_A_red_updist,col=col_Gfos_A, pch=16) points(Ar_Mod_B~meanAr_B_updist,col=col_Gfos_B, pch=16) for (j in 1:length(meanAr_A_red_updist)){ point <- cbind(meanAr_A_red_updist,Ar_Mod_A)[j,] seg <- unlist(perp.segment.coord(point[1],point[2])) segments(seg[1],seg[2],seg[3],seg[4], col=col_Gfos_A, lty=2, lwd=0.5) } for (k in 1:length(meanAr_B_updist)){ point <- cbind(meanAr_B_updist,Ar_Mod_B)[k,] seg <- unlist(perp.segment.coord(point[1],point[2])) segments(seg[1],seg[2],seg[3],seg[4], col=col_Gfos_B, lty=2, lwd=0.5) } text(max(Ar_modelled[,-1]),min(Ar_modelled[,-1])+1.5,paste0(measure1_short, " = ",formatC(round(sum(orthdist_Ar_A),sum_digits),digits=sum_digits, format="f")), adj=1, col=col_Gfos_A) text(max(Ar_modelled[,-1]),min(Ar_modelled[,-1])+0.5,paste0(measure1_short," = ",formatC(round(sum(orthdist_Ar_B),sum_digits),digits=sum_digits, format="f")), adj=1, col=col_Gfos_B) par(xpd=T) legend(-5,28,c(label_A,label_B),pch = 16, col = c(col_Gfos_A,col_Gfos_B), bty="n", cex=2) # text(0,0,"SOSO = Sum of squared orthogonals", adj=0) par(xpd=F) dev.off() ##** FIG 5 #### ##** Fst by instream distance (power function) by species, showing IBD if(fst){ if(pdf){ pdf(paste0(WD,"/Analysis_",output,"/Fig5.pdf"), width=8, height=6) }else{ png(paste0(WD,"/Analysis_",output,"/Fig5.png"), width=8, height=6, units="in", res=300) } GLMplot("fst","dist",dat=DISTDATA,model="slm_fst_power", CI_border = F, xlabel="Instream distance [km]",ylabel=expression('Genetic diff. [Pairwise Nei F'[ST]*']'),xax="n",pointtrans = T) axis(1, c(0,50,100,150,200,250), at=c(0,50000,100000,150000,200000,250000), cex.axis=1.5) dev.off() } #### SUPP INFO -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ #### dir.create(paste0(WD,"/Analysis_",output,"/SuppFigs"), showWarnings=F) ##** FIG S1 #### # create an overview map (once at the end of looping over parameter space) if(pdf){ pdf(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS1.pdf"), width=8, height=6) }else{ png(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS1.png"), width=8, height=6, units="in", res=300) } par(mar=c(0,0,0,0)) V(net)$modsite <- ifelse(V(net)$name%in%modsite,2,1) if (internal){ river_plot(overview_map = F, col_rhine=col_rhine, col_rhone=NA, col_ticino = NA, col_inn=NA, col_water = col_water, axes="none") }else{ river_plot(overview_map = F, col_rhine=col_rhine, col_rhone=NA, col_ticino = NA, col_inn=NA, col_water = col_water, axes="none") } plot(net, layout=net_layout, edge.arrow.size=0, edge.width=2.5, edge.color="blue", vertex.color=c(adjustcolor("white",0),"dark red")[V(net)$modsite], vertex.size=ifelse(V(net)$modsite==2,1,1), vertex.frame.color=col_water, vertex.shape="none", vertex.label=NA, rescale=F, xlim=c(min(net_layout[,1]), max(net_layout[,1])), ylim = c(min(net_layout[,2]), max(net_layout[,2])), asp = 0, add=T) # look at the existing/preloaded data cex_small=0.5 cex_big=0.75 if (internal){ points(Rhine$chx[Rhine$sample==0], Rhine$chy[Rhine$sample==0], pch=21, col="black", bg="white", cex=cex_small, lwd=0.75) # plot empty samples points(Rhine$chx[Rhine$sample==1], Rhine$chy[Rhine$sample==1], pch=21, col="black", bg="black", cex=cex_small, lwd=0.75) # plot amphipod data points(Gfos$chx[Gfos$type=="K"], Gfos$chy[Gfos$type=="K"], pch=21, col="black", bg=col_Gfos, cex=cex_big) # plot Gammarus fossarum complex points(Gfos$chx[Gfos$type=="A"], Gfos$chy[Gfos$type=="A"], pch=21, col="black", bg=col_Gfos_A, cex=cex_big) # plot Gammarus fossarum type A points(Gfos$chx[Gfos$type=="B"], Gfos$chy[Gfos$type=="B"], pch=21, col="black", bg=col_Gfos_B, cex=cex_big) # plot Gammarus fossarum type B points(Gfos$chx[Gfos$type=="C"], Gfos$chy[Gfos$type=="C"], pch=21, col="black", bg=col_Gfos, cex=cex_big) # plot Gammarus fossarum type C } points(site_coord$x, site_coord$y, col=c(1,1,0)[site_coord$gendata], bg=c(col_Gfos_A,col_Gfos_B,0)[site_coord$gendata], cex=1.5, pch=24) points(site_coord$x[site_coord$gendata==3], site_coord$y[site_coord$gendata==3], col=1, bg=col_Gfos_B, cex=1.5, pch=24) points(site_coord$x[site_coord$gendata==3], site_coord$y[site_coord$gendata==3], col=col_Gfos_A, cex=0.75, pch=17) text(465000,315000,"Microsat data (> 15 ind.)", adj=0) legend(465000,315000, c(label_A,label_B,"Both"), col=c(1,1,1), pt.bg=c(col_Gfos_A,col_Gfos_B,col_Gfos_B), pt.cex=1.5, pch=24, bty="n") legend(465000,315000, c("","",""), col=c(col_Gfos_A,col_Gfos_B,col_Gfos_A), pt.cex=0.75, pch=17, bty="n") text(465000,265000,"Presence/Absence data", adj=0, cex=0.8) legend(465000,265000, cex=0.8, c(expression(italic(G. ~ fossarum) ~ "complex"), label_A,label_B, "Amphipods present", "Amphipods absent"), col=c(1,1,1,1,1), pt.bg=c(col_Gfos,col_Gfos_A,col_Gfos_B,"black","white"), pt.cex=c(cex_big,cex_big,cex_big,cex_small,cex_small), pch=21, bty="n") dev.off() ##** FIG S2 #### ##** Correlation plot all explanatory variables if(pdf){ pdf(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS2.pdf"), width=8, height=6) }else{ png(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS2.png"), width=8, height=6, units="in", res=300) } PerformanceAnalytics::chart.Correlation(expl.var1, method = "kendall") dev.off() ##** FIG S3 #### ##** Correlation plot selected and transformed explanatory variables if(pdf){ pdf(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS3.pdf"), width=8, height=6) }else{ png(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS3.png"), width=8, height=6, units="in", res=300) } PerformanceAnalytics::chart.Correlation(expl.var, method = "kendall") dev.off() ##** FIG S4 #### #Perpendicular offset example x <- seq(1,10,1) y1 <- seq(1,10,1) y2 <- seq(3,12,1) y3 <- seq(-1,8,1) y4 <- seq(2,20,2) y5 <- seq(20,2,-2) y6 <- sample(seq(5.499,5.501,0.00001),10) y6 <- y6[order(y6)[c(1,10,3,8,5,6,7,4,9,2)]] cor1 <- cor(x,y1) cor2 <- cor(x,y2) cor3 <- cor(x,y3) cor4 <- cor(x,y4) cor5 <- cor(x,y5) cor6 <- abs(cor(x,y6)) orthdist_y1 <- c() orthdist_y1_directed <- c() orthdist_y2 <- c() orthdist_y2_directed <- c() orthdist_y3 <- c() orthdist_y3_directed <- c() orthdist_y4 <- c() orthdist_y4_directed <- c() orthdist_y5 <- c() orthdist_y5_directed <- c() orthdist_y6 <- c() orthdist_y6_directed <- c() # all combined in one plot col1 <- "darkgreen" col2 <- "steelblue" col3 <- "red" col4 <- "orange" col5 <- "purple" col6 <- "gold" plot(x,y1, xlim=c(min(x,y1,y2,y3,y4,y5,y6),max(x,y1,y2,y3,y4,y5,y6)), ylim=c(min(x,y1,y2,y3,y4,y5,y6),max(x,y1,y2,y3,y4,y5,y6)), asp=1, type="n", xlab="Empirical data", ylab="Simulated data") abline(0,1, lwd=1, lty=2) points(x,y1,col=col1, pch=16) for (i in 1:length(x)){ point <- cbind(x,y1)[i,] seg <- unlist(perp.segment.coord(point[1],point[2])) orthdist_y1[i] <- euc.dist(c(seg[1],seg[2]),c(seg[3],seg[4])) orthdist_y1_directed[i] <- sign(seg[2]-seg[4])*orthdist_y1[i] segments(seg[1],seg[2],seg[3],seg[4], col=col1, lty=2, lwd=0.5) } points(x,y2,col=col2, pch=16) for (i in 1:length(x)){ point <- cbind(x,y2)[i,] seg <- unlist(perp.segment.coord(point[1],point[2])) orthdist_y2[i] <- euc.dist(c(seg[1],seg[2]),c(seg[3],seg[4])) orthdist_y2_directed[i] <- sign(seg[2]-seg[4])*orthdist_y2[i] segments(seg[1],seg[2],seg[3],seg[4], col=col2, lty=2, lwd=0.5) } points(x,y3,col=col3, pch=16) for (i in 1:length(x)){ point <- cbind(x,y3)[i,] seg <- unlist(perp.segment.coord(point[1],point[2])) orthdist_y3[i] <- euc.dist(c(seg[1],seg[2]),c(seg[3],seg[4])) orthdist_y3_directed[i] <- sign(seg[2]-seg[4])*orthdist_y3[i] segments(seg[1],seg[2],seg[3],seg[4], col=col3, lty=2, lwd=0.5) } points(x,y4,col=col4, pch=16) for (i in 1:length(x)){ point <- cbind(x,y4)[i,] seg <- unlist(perp.segment.coord(point[1],point[2])) orthdist_y4[i] <- euc.dist(c(seg[1],seg[2]),c(seg[3],seg[4])) orthdist_y4_directed[i] <- sign(seg[2]-seg[4])*orthdist_y4[i] segments(seg[1],seg[2],seg[3],seg[4], col=col4, lty=2, lwd=0.5) } points(x,y5,col=col5, pch=16) for (i in 1:length(x)){ point <- cbind(x,y5)[i,] seg <- unlist(perp.segment.coord(point[1],point[2])) orthdist_y5[i] <- euc.dist(c(seg[1],seg[2]),c(seg[3],seg[4])) orthdist_y5_directed[i] <- sign(seg[2]-seg[4])*orthdist_y5[i] segments(seg[1],seg[2],seg[3],seg[4], col=col5, lty=2, lwd=0.5) } points(x,y6,col=col6, pch=16) for (i in 1:length(x)){ point <- cbind(x,y6)[i,] seg <- unlist(perp.segment.coord(point[1],point[2])) orthdist_y6[i] <- euc.dist(c(seg[1],seg[2]),c(seg[3],seg[4])) orthdist_y6_directed[i] <- sign(seg[2]-seg[4])*orthdist_y6[i] segments(seg[1],seg[2],seg[3],seg[4], col=col6, lty=2, lwd=0.5) } text(8,0,paste0("Cor: ", specify_decimal(cor1,1)), col=col1) text(15,0,paste0("SPO: ", specify_decimal(sum(orthdist_y1),1)), col=col1) text(20,0,paste0("MPO: ", specify_decimal(median(orthdist_y1),1)),col=col1) text(25,0,paste0("DMPO: ", specify_decimal(median(orthdist_y1_directed),1)), col=col1) text(8,1,paste0("Cor: ", specify_decimal(cor2,1)), col=col2) text(15,1,paste0("SPO: ", specify_decimal(sum(orthdist_y2),1)), col=col2) text(20,1,paste0("MPO: ", specify_decimal(median(orthdist_y2),1)),col=col2) text(25,1,paste0("DMPO: ", specify_decimal(median(orthdist_y2_directed),1)), col=col2) text(8,2,paste0("Cor: ", specify_decimal(cor3,1)), col=col3) text(15,2,paste0("SPO: ", specify_decimal(sum(orthdist_y3),1)), col=col3) text(20,2,paste0("MPO: ", specify_decimal(median(orthdist_y3),1)),col=col3) text(25,2,paste0("DMPO: ", specify_decimal(median(orthdist_y3_directed),1)), col=col3) text(8,3,paste0("Cor: ", specify_decimal(cor4,1)), col=col4) text(15,3,paste0("SPO: ", specify_decimal(sum(orthdist_y4),1)), col=col4) text(20,3,paste0("MPO: ", specify_decimal(median(orthdist_y4),1)),col=col4) text(25,3,paste0("DMPO: ", specify_decimal(median(orthdist_y4_directed),1)), col=col4) text(8,4,paste0("Cor: ", specify_decimal(cor5,1)), col=col5) text(15,4,paste0("SPO: ", specify_decimal(sum(orthdist_y5),1)), col=col5) text(20,4,paste0("MPO: ", specify_decimal(median(orthdist_y5),1)),col=col5) text(25,4,paste0("DMPO: ", specify_decimal(median(orthdist_y5_directed),1)), col=col5) text(8,5,paste0("Cor: ", specify_decimal(cor6,1)), col=col6) text(15,5,paste0("SPO: ", specify_decimal(sum(orthdist_y6),1)), col=col6) text(20,5,paste0("MPO: ", specify_decimal(median(orthdist_y6),1)),col=col6) text(25,5,paste0("DMPO: ", specify_decimal(median(orthdist_y6_directed),1)), col=col6) # six plots if(pdf){ pdf(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS4.pdf"), width=9, height=6) }else{ png(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS4.png"), width=9, height=6, units="in", res=300) } nf <- multipanel.layout(main.col=1,main.row=1,sub.col=3,sub.row=2,sub1=F,sub2=F,main=pla$main, w.legend=0.05, w.axis=0.3, h.axis=0.3) x <- seq(1,10,1) y1 <- seq(1,10,1) y2 <- seq(3,12,1) y3 <- seq(-1,8,1) y4 <- seq(2,20,2) y5 <- seq(20,2,-2) while(cor6>=0.05){ y6 <- sample(seq(5.499,5.501,0.00001),10) y6 <- y6[order(y6)[c(1,10,3,8,5,6,7,4,9,2)]] cor6 <- abs(cor(x,y6)) } par(mar=c(0,0,0,0)) x_text <- 20 y_text_1 <- 6 y_text_2 <- 4 y_text_3 <- 2 y_text_4 <- 0 cex_perp <- 2 col_perp <- "steelblue" col1 <- col_perp col2 <- col_perp col3 <- col_perp col4 <- col_perp col5 <- col_perp col6 <- col_perp lwd_perp <- 1 pt.cex <- 2 plot(x,y1, xlim=c(min(x,y1,y2,y3,y4,y5,y6),max(x,y1,y2,y3,y4,y5,y6)), ylim=c(min(x,y1,y2,y3,y4,y5,y6),max(x,y1,y2,y3,y4,y5,y6)), asp=1, type="n", xlab="Empirical data", ylab="Simulated data", xaxt="n", yaxt="n") abline(0,1, lwd=1, lty=2) points(x,y1,col=col1, pch=16, cex=pt.cex) for (i in 1:length(x)){ point <- cbind(x,y1)[i,] seg <- unlist(perp.segment.coord(point[1],point[2])) orthdist_y1[i] <- euc.dist(c(seg[1],seg[2]),c(seg[3],seg[4])) orthdist_y1_directed[i] <- sign(seg[2]-seg[4])*orthdist_y1[i] segments(seg[1],seg[2],seg[3],seg[4], col=col1, lty=2, lwd=lwd_perp) } text(par("usr")[1]+((par("usr")[2]-par("usr")[1])/40),par("usr")[4]-((par("usr")[4]-par("usr")[3])/40), "(a)", adj=c(0,1), cex=cex_perp) text(x_text,y_text_1,paste0("Cor: ", specify_decimal(cor1,1)), col=col1, adj=1, cex=cex_perp) text(x_text,y_text_2,paste0("SPO: ", specify_decimal(sum(orthdist_y1),1)), col=col1, adj=1, cex=cex_perp) text(x_text,y_text_3,paste0("MPO: ", specify_decimal(median(orthdist_y1),1)),col=col1, adj=1, cex=cex_perp) text(x_text,y_text_4,paste0("DMPO: ", specify_decimal(median(orthdist_y1_directed),1)), col=col1, adj=1, cex=cex_perp) axis(2) plot(x,y4, xlim=c(min(x,y1,y2,y3,y4,y5,y6),max(x,y1,y2,y3,y4,y5,y6)), ylim=c(min(x,y1,y2,y3,y4,y5,y6),max(x,y1,y2,y3,y4,y5,y6)), asp=1, type="n", xlab="Empirical data", ylab="Simulated data", xaxt="n", yaxt="n") abline(0,1, lwd=1, lty=2) points(x,y4,col=col4, pch=16, cex=pt.cex) for (i in 1:length(x)){ point <- cbind(x,y4)[i,] seg <- unlist(perp.segment.coord(point[1],point[2])) orthdist_y4[i] <- euc.dist(c(seg[1],seg[2]),c(seg[3],seg[4])) orthdist_y4_directed[i] <- sign(seg[2]-seg[4])*orthdist_y4[i] segments(seg[1],seg[2],seg[3],seg[4], col=col4, lty=2, lwd=lwd_perp) } text(par("usr")[1]+((par("usr")[2]-par("usr")[1])/40),par("usr")[4]-((par("usr")[4]-par("usr")[3])/40), "(d)", adj=c(0,1), cex=cex_perp) text(x_text,y_text_1,paste0("Cor: ", specify_decimal(cor4,1)), col=col4, adj=1, cex=cex_perp) text(x_text,y_text_2,paste0("SPO: ", specify_decimal(sum(orthdist_y4),1)), col=col4, adj=1, cex=cex_perp) text(x_text,y_text_3,paste0("MPO: ", specify_decimal(median(orthdist_y4),1)),col=col4, adj=1, cex=cex_perp) text(x_text,y_text_4,paste0("DMPO: ", specify_decimal(median(orthdist_y4_directed),1)), col=col4, adj=1, cex=cex_perp) axis(1) axis(2) plot(x,y2, xlim=c(min(x,y1,y2,y3,y4,y5,y6),max(x,y1,y2,y3,y4,y5,y6)), ylim=c(min(x,y1,y2,y3,y4,y5,y6),max(x,y1,y2,y3,y4,y5,y6)), asp=1, type="n", xlab="Empirical data", ylab="Simulated data", xaxt="n", yaxt="n") abline(0,1, lwd=1, lty=2) points(x,y2,col=col2, pch=16, cex=pt.cex) for (i in 1:length(x)){ point <- cbind(x,y2)[i,] seg <- unlist(perp.segment.coord(point[1],point[2])) orthdist_y2[i] <- euc.dist(c(seg[1],seg[2]),c(seg[3],seg[4])) orthdist_y2_directed[i] <- sign(seg[2]-seg[4])*orthdist_y2[i] segments(seg[1],seg[2],seg[3],seg[4], col=col2, lty=2, lwd=lwd_perp) } text(par("usr")[1]+((par("usr")[2]-par("usr")[1])/40),par("usr")[4]-((par("usr")[4]-par("usr")[3])/40), "(b)", adj=c(0,1), cex=cex_perp) text(x_text,y_text_1,paste0("Cor: ", specify_decimal(cor2,1)), col=col2, adj=1, cex=cex_perp) text(x_text,y_text_2,paste0("SPO: ", specify_decimal(sum(orthdist_y2),1)), col=col2, adj=1, cex=cex_perp) text(x_text,y_text_3,paste0("MPO: ", specify_decimal(median(orthdist_y2),1)),col=col2, adj=1, cex=cex_perp) text(x_text,y_text_4,paste0("DMPO: ", specify_decimal(median(orthdist_y2_directed),1)), col=col2, adj=1, cex=cex_perp) plot(x,y5, xlim=c(min(x,y1,y2,y3,y4,y5,y6),max(x,y1,y2,y3,y4,y5,y6)), ylim=c(min(x,y1,y2,y3,y4,y5,y6),max(x,y1,y2,y3,y4,y5,y6)), asp=1, type="n", xlab="Empirical data", ylab="Simulated data", xaxt="n", yaxt="n") abline(0,1, lwd=1, lty=2) points(x,y5,col=col5, pch=16, cex=pt.cex) for (i in 1:length(x)){ point <- cbind(x,y5)[i,] seg <- unlist(perp.segment.coord(point[1],point[2])) orthdist_y5[i] <- euc.dist(c(seg[1],seg[2]),c(seg[3],seg[4])) orthdist_y5_directed[i] <- sign(seg[2]-seg[4])*orthdist_y5[i] segments(seg[1],seg[2],seg[3],seg[4], col=col5, lty=2, lwd=lwd_perp) } text(par("usr")[1]+((par("usr")[2]-par("usr")[1])/40),par("usr")[4]-((par("usr")[4]-par("usr")[3])/40), "(e)", adj=c(0,1), cex=cex_perp) text(x_text,y_text_1,paste0("Cor: ", specify_decimal(cor5,1)), col=col5, adj=1, cex=cex_perp) text(x_text,y_text_2,paste0("SPO: ", specify_decimal(sum(orthdist_y5),1)), col=col5, adj=1, cex=cex_perp) text(x_text,y_text_3,paste0("MPO: ", specify_decimal(median(orthdist_y5),1)),col=col5, adj=1, cex=cex_perp) text(x_text,y_text_4,paste0("DMPO: ", specify_decimal(median(orthdist_y5_directed),1)), col=col5, adj=1, cex=cex_perp) axis(1) plot(x,y3, xlim=c(min(x,y1,y2,y3,y4,y5,y6),max(x,y1,y2,y3,y4,y5,y6)), ylim=c(min(x,y1,y2,y3,y4,y5,y6),max(x,y1,y2,y3,y4,y5,y6)), asp=1, type="n", xlab="Empirical data", ylab="Simulated data", xaxt="n", yaxt="n") abline(0,1, lwd=1, lty=2) points(x,y3,col=col3, pch=16, cex=pt.cex) for (i in 1:length(x)){ point <- cbind(x,y3)[i,] seg <- unlist(perp.segment.coord(point[1],point[2])) orthdist_y3[i] <- euc.dist(c(seg[1],seg[2]),c(seg[3],seg[4])) orthdist_y3_directed[i] <- sign(seg[2]-seg[4])*orthdist_y3[i] segments(seg[1],seg[2],seg[3],seg[4], col=col3, lty=2, lwd=lwd_perp) } text(par("usr")[1]+((par("usr")[2]-par("usr")[1])/40),par("usr")[4]-((par("usr")[4]-par("usr")[3])/40), "(c)", adj=c(0,1), cex=cex_perp) text(x_text,y_text_1,paste0("Cor: ", specify_decimal(cor3,1)), col=col3, adj=1, cex=cex_perp) text(x_text,y_text_2,paste0("SPO: ", specify_decimal(sum(orthdist_y3),1)), col=col3, adj=1, cex=cex_perp) text(x_text,y_text_3,paste0("MPO: ", specify_decimal(median(orthdist_y3),1)),col=col3, adj=1, cex=cex_perp) text(x_text,y_text_4,paste0("DMPO: ", specify_decimal(median(orthdist_y3_directed),1)), col=col3, adj=1, cex=cex_perp) plot(x,y6, xlim=c(min(x,y1,y2,y3,y4,y5,y6),max(x,y1,y2,y3,y4,y5,y6)), ylim=c(min(x,y1,y2,y3,y4,y5,y6),max(x,y1,y2,y3,y4,y5,y6)), asp=1, type="n", xlab="Empirical data", ylab="Simulated data", xaxt="n", yaxt="n") abline(0,1, lwd=1, lty=2) points(x,y6,col=col6, pch=16, cex=pt.cex) for (i in 1:length(x)){ point <- cbind(x,y6)[i,] seg <- unlist(perp.segment.coord(point[1],point[2])) orthdist_y6[i] <- euc.dist(c(seg[1],seg[2]),c(seg[3],seg[4])) orthdist_y6_directed[i] <- sign(seg[2]-seg[4])*orthdist_y6[i] segments(seg[1],seg[2],seg[3],seg[4], col=col6, lty=2, lwd=lwd_perp) } text(par("usr")[1]+((par("usr")[2]-par("usr")[1])/40),par("usr")[4]-((par("usr")[4]-par("usr")[3])/40), "(f)", adj=c(0,1), cex=cex_perp) text(x_text,y_text_1,paste0("Cor: ", specify_decimal(cor6,1)), col=col6, adj=1, cex=cex_perp) text(x_text,y_text_2,paste0("SPO: ", specify_decimal(sum(orthdist_y6),1)), col=col6, adj=1, cex=cex_perp) text(x_text,y_text_3,paste0("MPO: ", specify_decimal(median(orthdist_y6),1)),col=col6, adj=1, cex=cex_perp) text(x_text,y_text_4,paste0("DMPO: ", specify_decimal(median(orthdist_y6_directed),1)), col=col6, adj=1, cex=cex_perp) axis(1) plot.new() text(0.5,0.5,"Simulated data", cex=2, srt = 90) plot.new() text(0.5,0.5,"Simulated data", cex=2, srt = 90) plot.new() text(0.5,0.5,"Empirical data", cex=2) plot.new() text(0.5,0.5,"Empirical data", cex=2) plot.new() text(0.5,0.5,"Empirical data", cex=2) dev.off() ##** FIG S5 #### #*** Mean Ho maps mp_dim <- multipanel.dimensions(main.col=pla$a,main.row=pla$b,sub.row=pla$x,sub.col=pla$y,sub1=pla$sub1,sub2=pla$sub2, main=T, h.main=0.5, w.legend=0, h.sub2=0.3, w.axis=0.7, h.axis=0, spacer.sub.col=0.5) if(pdf){ pdf(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS5.pdf"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1])) }else{ png(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS5.png"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1]), units="in", res=600) } par(mar=c(0,0,0,0)) nf <- multipanel.layout(main.col=pla$a,main.row=pla$b,sub.row=pla$x,sub.col=pla$y,sub1=pla$sub1,sub2=pla$sub2, main=T, h.main=0.5, w.legend=0, h.sub2=0.3, w.axis=0.7, h.axis=0, spacer.sub.col=0.5) for (j in 2:ncol(Ho_modelled)){ if (internal){ river_plot(width_country=0.5, lwd_rivers=0.5, xlimit=c(495000,825000), col_water=col_water, north_arrow = F, overview_map = F, scalebar=F, arrows = F, border_outline=F, width_border=2, col_rhine = col_rhine, plot_rhone=F, lines_rhone=F, plot_ticino=F, lines_ticino=F, plot_inn=F, lines_inn=F, lakes=TRUE, rivers=TRUE, axes="none", river_nr=T) }else{ river_plot(width_country=0.5, lwd_rivers=0.5, xlimit=c(495000,825000), col_water=col_water, north_arrow = F, overview_map = F, scalebar=F, arrows = F, border_outline=F, width_border=2, col_rhine = col_rhine, plot_rhone=F, plot_ticino=F, plot_inn=F, lakes=TRUE, rivers=TRUE, axes="none", river_nr=F) } values <- (Ho_modelled[,j]-min(Ho_modelled[,j]))/ (max(Ho_modelled[,j])-min(Ho_modelled[,j])) # transform meanAr values to range [0,1] for heatmap plotting if(!any(is.na(values))){ for(i in 1:length(match_Mod)){ points(site_coord$x[match_Mod[i]], site_coord$y[match_Mod[i]], bg=rgb2hex(col_fun(values))[i], pch=21, lwd=0.5, cex=1) } } if (j %in% c(2,3)){ mtext(lab[6], side=3, line=-1, cex=1.3) } if (j %in% c(8,9)){ mtext(lab[7], side=3, line=-1, cex=1.3) } if (j %in% c(14,15)){ mtext(lab[8], side=3, line=-1, cex=1.3) } gradient.legend(Ho_modelled[,j],alpha=1, val.midpoint=F, round=2, val.cex=1, val.gap = 0.5, title.gap=0.1, xl = 505000, xr = 635000, yb = 20000, yt = 40000, horizontal=T) } plot.new() text(0.5,0.5,"Simulated observed heterozygosity",adj=c(0.5,0.5), cex=3) for (i in 1:5){ plot.new() if (i %in% c(1:2)){ text(0.5,1,lab[i],adj=c(0.5,1), cex=lab_cex[i]) } if (i %in% c(3:5)){ text(0.5,0.5,lab[i],adj=c(0.5,0.5), cex=lab_cex[i]) } } dev.off() ##** FIG S6 #### #*** Mean Hs maps mp_dim <- multipanel.dimensions(main.col=pla$a,main.row=pla$b,sub.row=pla$x,sub.col=pla$y,sub1=pla$sub1,sub2=pla$sub2, main=T, h.main=0.5, w.legend=0, h.sub2=0.3, w.axis=0.7, h.axis=0, spacer.sub.col=0.5) if(pdf){ pdf(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS6.pdf"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1])) }else{ png(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS6.png"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1]), units="in", res=600) } par(mar=c(0,0,0,0)) nf <- multipanel.layout(main.col=pla$a,main.row=pla$b,sub.row=pla$x,sub.col=pla$y,sub1=pla$sub1,sub2=pla$sub2, main=T, h.main=0.5, w.legend=0, h.sub2=0.3, w.axis=0.7, h.axis=0, spacer.sub.col=0.5) for (j in 2:ncol(Hs_modelled)){ if (internal){ river_plot(width_country=0.5, lwd_rivers=0.5, xlimit=c(495000,825000), col_water=col_water, north_arrow = F, overview_map = F, scalebar=F, arrows = F, border_outline=F, width_border=2, col_rhine = col_rhine, plot_rhone=F, lines_rhone=F, plot_ticino=F, lines_ticino=F, plot_inn=F, lines_inn=F, lakes=TRUE, rivers=TRUE, axes="none", river_nr=T) }else{ river_plot(width_country=0.5, lwd_rivers=0.5, xlimit=c(495000,825000), col_water=col_water, north_arrow = F, overview_map = F, scalebar=F, arrows = F, border_outline=F, width_border=2, col_rhine = col_rhine, plot_rhone=F, plot_ticino=F, plot_inn=F, lakes=TRUE, rivers=TRUE, axes="none", river_nr=F) } values <- (Hs_modelled[,j]-min(Hs_modelled[,j]))/ (max(Hs_modelled[,j])-min(Hs_modelled[,j])) # transform meanAr values to range [0,1] for heatmap plotting if(!any(is.na(values))){ for(i in 1:length(match_Mod)){ points(site_coord$x[match_Mod[i]], site_coord$y[match_Mod[i]], bg=rgb2hex(col_fun(values))[i], pch=21, lwd=0.5, cex=1) } } if (j %in% c(2,3)){ mtext(lab[6], side=3, line=-1, cex=1.3) } if (j %in% c(8,9)){ mtext(lab[7], side=3, line=-1, cex=1.3) } if (j %in% c(14,15)){ mtext(lab[8], side=3, line=-1, cex=1.3) } gradient.legend(Hs_modelled[,j],alpha=1, val.midpoint=F, round=2, val.cex=1, val.gap = 0.5, title.gap=0.1, xl = 505000, xr = 635000, yb = 20000, yt = 40000, horizontal=T) } plot.new() text(0.5,0.5,"Simulated expected heterozygosity",adj=c(0.5,0.5), cex=3) for (i in 1:5){ plot.new() if (i %in% c(1:2)){ text(0.5,1,lab[i],adj=c(0.5,1), cex=lab_cex[i]) } if (i %in% c(3:5)){ text(0.5,0.5,lab[i],adj=c(0.5,0.5), cex=lab_cex[i]) } } dev.off() ##** W: Model performance #### maxhist_Ar <- ceiling(max(hist_Ar_A,hist_Ar_B, na.rm=T)) maxhist_Ho <- ceiling(10*max(hist_Ho_A,hist_Ho_B, na.rm=T))/10 maxhist_Hs <- ceiling(10*max(hist_Hs_A,hist_Hs_B, na.rm=T))/10 w00_Ar_A <- hist_Ar_A[,c(1,2,7,8,13,14)] w05_Ar_A <- hist_Ar_A[,c(3,4,9,10,15,16)] w10_Ar_A <- hist_Ar_A[,c(5,6,11,12,17,18)] w00_Ar_B <- hist_Ar_B[,c(1,2,7,8,13,14)] w05_Ar_B <- hist_Ar_B[,c(3,4,9,10,15,16)] w10_Ar_B <- hist_Ar_B[,c(5,6,11,12,17,18)] w00_Ho_A <- hist_Ho_A[,c(1,2,7,8,13,14)] w05_Ho_A <- hist_Ho_A[,c(3,4,9,10,15,16)] w10_Ho_A <- hist_Ho_A[,c(5,6,11,12,17,18)] w00_Ho_B <- hist_Ho_B[,c(1,2,7,8,13,14)] w05_Ho_B <- hist_Ho_B[,c(3,4,9,10,15,16)] w10_Ho_B <- hist_Ho_B[,c(5,6,11,12,17,18)] w00_Hs_A <- hist_Hs_A[,c(1,2,7,8,13,14)] w05_Hs_A <- hist_Hs_A[,c(3,4,9,10,15,16)] w10_Hs_A <- hist_Hs_A[,c(5,6,11,12,17,18)] w00_Hs_B <- hist_Hs_B[,c(1,2,7,8,13,14)] w05_Hs_B <- hist_Hs_B[,c(3,4,9,10,15,16)] w10_Hs_B <- hist_Hs_B[,c(5,6,11,12,17,18)] diffSPO_w00both_Ar_A <- c(apply(w00_Ar_A,2,sum)-apply(w05_Ar_A,2,sum),apply(w00_Ar_A,2,sum)-apply(w10_Ar_A,2,sum)) diffSPO_w00both_Ar_B <- c(apply(w00_Ar_B,2,sum)-apply(w05_Ar_B,2,sum),apply(w00_Ar_B,2,sum)-apply(w10_Ar_B,2,sum)) diffSPO_w00both_Ho_A <- c(apply(w00_Ho_A,2,sum)-apply(w05_Ho_A,2,sum),apply(w00_Ho_A,2,sum)-apply(w10_Ho_A,2,sum)) diffSPO_w00both_Ho_B <- c(apply(w00_Ho_B,2,sum)-apply(w05_Ho_B,2,sum),apply(w00_Ho_B,2,sum)-apply(w10_Ho_B,2,sum)) diffSPO_w00both_Hs_A <- c(apply(w00_Hs_A,2,sum)-apply(w05_Hs_A,2,sum),apply(w00_Hs_A,2,sum)-apply(w10_Hs_A,2,sum)) diffSPO_w00both_Hs_B <- c(apply(w00_Hs_B,2,sum)-apply(w05_Hs_B,2,sum),apply(w00_Hs_B,2,sum)-apply(w10_Hs_B,2,sum)) diffMPO_w00both_Ar_A <- c(apply(w00_Ar_A,2,median)-apply(w05_Ar_A,2,median),apply(w00_Ar_A,2,median)-apply(w10_Ar_A,2,median)) diffMPO_w00both_Ar_B <- c(apply(w00_Ar_B,2,median)-apply(w05_Ar_B,2,median),apply(w00_Ar_B,2,median)-apply(w10_Ar_B,2,median)) diffMPO_w00both_Ho_A <- c(apply(w00_Ho_A,2,median)-apply(w05_Ho_A,2,median),apply(w00_Ho_A,2,median)-apply(w10_Ho_A,2,median)) diffMPO_w00both_Ho_B <- c(apply(w00_Ho_B,2,median)-apply(w05_Ho_B,2,median),apply(w00_Ho_B,2,median)-apply(w10_Ho_B,2,median)) diffMPO_w00both_Hs_A <- c(apply(w00_Hs_A,2,median)-apply(w05_Hs_A,2,median),apply(w00_Hs_A,2,median)-apply(w10_Hs_A,2,median)) diffMPO_w00both_Hs_B <- c(apply(w00_Hs_B,2,median)-apply(w05_Hs_B,2,median),apply(w00_Hs_B,2,median)-apply(w10_Hs_B,2,median)) list_w00both <- c(diffMPO_w00both_Ar_A,diffMPO_w00both_Ar_B, diffMPO_w00both_Ho_A,diffMPO_w00both_Ho_B, diffMPO_w00both_Hs_A,diffMPO_w00both_Hs_B, diffSPO_w00both_Ar_A,diffSPO_w00both_Ar_B, diffSPO_w00both_Ho_A,diffSPO_w00both_Ho_B, diffSPO_w00both_Hs_A,diffSPO_w00both_Hs_B) total_comparison_w00both <- length(list_w00both) improved_fit_w00both <- sum(list_w00both<0) # Model performance W improved_fit_w00both/total_comparison_w00both ##** FIG S7 #### ##** Perpendicular offset histogram comparison for W if(pdf){ pdf(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS7.pdf"), width=8, height=6) }else{ png(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS7.png"), width=8, height=6, units="in", res=300) } op <- par(mfrow = c(3,3), oma = c(5,5,2,0) + 0.1, mar = c(0,0,2,1) + 0.1) hist(w00_Ar_A, col=col_Gfos_A, cex.axis=2, xaxt="n", main="", ylim=c(0,70), xlim = c(0,maxhist_Ar), breaks=seq(0,maxhist_Ar,0.25)) hist(w00_Ar_B, col=col_Gfos_B, add=T, breaks=seq(0,maxhist_Ar,0.25)) axis(1, labels=F, cex.axis=2) mtext(side = 3, text = lab_Ar_short, line = 1, adj=0.5, cex = 1.5) abline(v=median(w00_Ar_A, na.rm=T), col=white_transparent, lwd=2) abline(v=median(w00_Ar_A, na.rm=T), col=col_Gfos_A, lty=2, lwd=2) abline(v=median(w00_Ar_B, na.rm=T), col=white_transparent, lwd=2) abline(v=median(w00_Ar_B, na.rm=T), col=col_Gfos_B, lty=2, lwd=2) textbox(4,65, txt = lab[3], txt.adj=0.5, txt.cex = 2, frm.brd = NA, frm.col = white_transparent) legend(2,50, c(expression("Median" ~ italic(G. ~ fossarum) ~ "type A"),expression("Median" ~ italic(G. ~ fossarum) ~ "type B")), lty=2, col=c(col_Gfos_A,col_Gfos_B), bty="n", cex=0.85, lwd=2) hist(w00_Ho_A, col=col_Gfos_A, yaxt="n", xaxt="n", main="", ylim=c(0,70), xlim = c(0,maxhist_Ho), breaks=seq(0,maxhist_Ho,0.025)) hist(w00_Ho_B, col=col_Gfos_B, add=T, breaks=seq(0,maxhist_Ho,0.025)) axis(1, labels=F, cex.axis=2) axis(2, labels=F, cex.axis=2) mtext(side = 3, text = lab_Ho_short, line = 1, adj=0.5, cex = 1.5) abline(v=median(w00_Ho_A, na.rm=T), col=white_transparent, lwd=2) abline(v=median(w00_Ho_A, na.rm=T), col=col_Gfos_A, lty=2, lwd=2) abline(v=median(w00_Ho_B, na.rm=T), col=white_transparent, lwd=2) abline(v=median(w00_Ho_B, na.rm=T), col=col_Gfos_B, lty=2, lwd=2) textbox(0.3,65, txt = lab[3], txt.adj=0.5, txt.cex = 2, frm.brd = NA, frm.col = white_transparent) hist(w00_Hs_A, col=col_Gfos_A, yaxt="n", xaxt="n", main="", ylim=c(0,70), xlim = c(0,maxhist_Hs), breaks=seq(0,maxhist_Hs,0.025)) hist(w00_Hs_B, col=col_Gfos_B, add=T, breaks=seq(0,maxhist_Hs,0.025)) axis(1, labels=F, cex.axis=2) axis(2, labels=F, cex.axis=2) mtext(side = 3, text = lab_Hs_short, line = 1, adj=0.5, cex = 1.5) abline(v=median(w00_Hs_A, na.rm=T), col=white_transparent, lwd=2) abline(v=median(w00_Hs_A, na.rm=T), col=col_Gfos_A, lty=2, lwd=2) abline(v=median(w00_Hs_B, na.rm=T), col=white_transparent, lwd=2) abline(v=median(w00_Hs_B, na.rm=T), col=col_Gfos_B, lty=2, lwd=2) textbox(0.3,65, txt = lab[3], txt.adj=0.5, txt.cex = 2, frm.brd = NA, frm.col = white_transparent) hist(w05_Ar_A, col=col_Gfos_A, cex.axis=2, xaxt="n", main="", ylim=c(0,70), xlim = c(0,maxhist_Ar), breaks=seq(0,maxhist_Ar,0.25)) hist(w05_Ar_B, col=col_Gfos_B, add=T, breaks=seq(0,maxhist_Ar,0.25)) axis(1, labels=F, cex.axis=2) abline(v=median(w05_Ar_A, na.rm=T), col=white_transparent, lwd=2) abline(v=median(w05_Ar_A, na.rm=T), col=col_Gfos_A, lty=2, lwd=2) abline(v=median(w05_Ar_B, na.rm=T), col=white_transparent, lwd=2) abline(v=median(w05_Ar_B, na.rm=T), col=col_Gfos_B, lty=2, lwd=2) textbox(4,65, txt = lab[4], txt.adj=0.5, txt.cex = 2, frm.brd = NA, frm.col = white_transparent) hist(w05_Ho_A, col=col_Gfos_A, yaxt="n", xaxt="n", main="", ylim=c(0,70), xlim = c(0,maxhist_Ho), breaks=seq(0,maxhist_Ho,0.025)) hist(w05_Ho_B, col=col_Gfos_B, add=T, breaks=seq(0,maxhist_Ho,0.025)) axis(1, labels=F, cex.axis=2) axis(2, labels=F, cex.axis=2) abline(v=median(w05_Ho_A, na.rm=T), col=white_transparent, lwd=2) abline(v=median(w05_Ho_A, na.rm=T), col=col_Gfos_A, lty=2, lwd=2) abline(v=median(w05_Ho_B, na.rm=T), col=white_transparent, lwd=2) abline(v=median(w05_Ho_B, na.rm=T), col=col_Gfos_B, lty=2, lwd=2) textbox(0.3,65, txt = lab[4], txt.adj=0.5, txt.cex = 2, frm.brd = NA, frm.col = white_transparent) hist(w05_Hs_A, col=col_Gfos_A, yaxt="n", xaxt="n", main="", ylim=c(0,70), xlim = c(0,maxhist_Hs), breaks=seq(0,maxhist_Hs,0.025)) hist(w05_Hs_B, col=col_Gfos_B, add=T, breaks=seq(0,maxhist_Hs,0.025)) axis(1, labels=F, cex.axis=2) axis(2, labels=F, cex.axis=2) abline(v=median(w05_Hs_A, na.rm=T), col=white_transparent, lwd=2) abline(v=median(w05_Hs_A, na.rm=T), col=col_Gfos_A, lty=2, lwd=2) abline(v=median(w05_Hs_B, na.rm=T), col=white_transparent, lwd=2) abline(v=median(w05_Hs_B, na.rm=T), col=col_Gfos_B, lty=2, lwd=2) textbox(0.3,65, txt = lab[4], txt.adj=0.5, txt.cex = 2, frm.brd = NA, frm.col = white_transparent) hist(w10_Ar_A, col=col_Gfos_A, cex.axis=2, xaxt="n", main="", ylim=c(0,70), xlim = c(0,maxhist_Ar), breaks=seq(0,maxhist_Ar,0.25)) hist(w10_Ar_B, col=col_Gfos_B, add=T, breaks=seq(0,maxhist_Ar,0.25)) axis(1, labels=T, cex.axis=2, padj=0.5) abline(v=median(w10_Ar_A, na.rm=T), col=white_transparent, lwd=2) abline(v=median(w10_Ar_A, na.rm=T), col=col_Gfos_A, lty=2, lwd=2) abline(v=median(w10_Ar_B, na.rm=T), col=white_transparent, lwd=2) abline(v=median(w10_Ar_B, na.rm=T), col=col_Gfos_B, lty=2, lwd=2) textbox(4,65, txt = lab[5], txt.adj=0.5, txt.cex = 2, frm.brd = NA, frm.col = white_transparent) hist(w10_Ho_A, col=col_Gfos_A, yaxt="n", xaxt="n", main="", ylim=c(0,70), xlim = c(0,maxhist_Ho), breaks=seq(0,maxhist_Ho,0.025)) hist(w10_Ho_B, col=col_Gfos_B, add=T, breaks=seq(0,maxhist_Ho,0.025)) axis(1, labels=T, cex.axis=2, padj=0.5) axis(2, labels=F, cex.axis=2) abline(v=median(w10_Ho_A, na.rm=T), col=white_transparent, lwd=2) abline(v=median(w10_Ho_A, na.rm=T), col=col_Gfos_A, lty=2, lwd=2) abline(v=median(w10_Ho_B, na.rm=T), col=white_transparent, lwd=2) abline(v=median(w10_Ho_B, na.rm=T), col=col_Gfos_B, lty=2, lwd=2) textbox(0.3,65, txt = lab[5], txt.adj=0.5, txt.cex = 2, frm.brd = NA, frm.col = white_transparent) hist(w10_Hs_A, col=col_Gfos_A, yaxt="n", xaxt="n", main="", ylim=c(0,70), xlim = c(0,maxhist_Hs), breaks=seq(0,maxhist_Hs,0.025)) hist(w10_Hs_B, col=col_Gfos_B, add=T, breaks=seq(0,maxhist_Hs,0.025)) axis(1, labels=T, cex.axis=2, padj=0.5) axis(2, labels=F, cex.axis=2) abline(v=median(w10_Hs_A, na.rm=T), col=white_transparent, lwd=2) abline(v=median(w10_Hs_A, na.rm=T), col=col_Gfos_A, lty=2, lwd=2) abline(v=median(w10_Hs_B, na.rm=T), col=white_transparent, lwd=2) abline(v=median(w10_Hs_B, na.rm=T), col=col_Gfos_B, lty=2, lwd=2) textbox(0.3,65, txt = lab[5], txt.adj=0.5, txt.cex = 2, frm.brd = NA, frm.col = white_transparent) strwidth(lab[5]) * 2 title(xlab = "Perpendicular offset", ylab = "Frequency", outer = TRUE, line = 3.5, cex.lab=2) dev.off() ##** d: Model performance #### d0001_Ar_A <- hist_Ar_A[,(1:6)] d001_Ar_A <- hist_Ar_A[,c(7:12)] d01_Ar_A <- hist_Ar_A[,c(13:18)] d0001_Ar_B <- hist_Ar_B[,(1:6)] d001_Ar_B <- hist_Ar_B[,c(7:12)] d01_Ar_B <- hist_Ar_B[,c(13:18)] d0001_Ho_A <- hist_Ho_A[,(1:6)] d001_Ho_A <- hist_Ho_A[,c(7:12)] d01_Ho_A <- hist_Ho_A[,c(13:18)] d0001_Ho_B <- hist_Ho_B[,(1:6)] d001_Ho_B <- hist_Ho_B[,c(7:12)] d01_Ho_B <- hist_Ho_B[,c(13:18)] d0001_Hs_A <- hist_Hs_A[,(1:6)] d001_Hs_A <- hist_Hs_A[,c(7:12)] d01_Hs_A <- hist_Hs_A[,c(13:18)] d0001_Hs_B <- hist_Hs_B[,(1:6)] d001_Hs_B <- hist_Hs_B[,c(7:12)] d01_Hs_B <- hist_Hs_B[,c(13:18)] diffSPO_d0001both_Ar_A <- c(apply(d0001_Ar_A,2,sum)-apply(d001_Ar_A,2,sum),apply(d0001_Ar_A,2,sum)-apply(d01_Ar_A,2,sum)) diffSPO_d0001both_Ar_B <- c(apply(d0001_Ar_B,2,sum)-apply(d001_Ar_B,2,sum),apply(d0001_Ar_B,2,sum)-apply(d01_Ar_B,2,sum)) diffSPO_d0001both_Ho_A <- c(apply(d0001_Ho_A,2,sum)-apply(d001_Ho_A,2,sum),apply(d0001_Ho_A,2,sum)-apply(d01_Ho_A,2,sum)) diffSPO_d0001both_Ho_B <- c(apply(d0001_Ho_B,2,sum)-apply(d001_Ho_B,2,sum),apply(d0001_Ho_B,2,sum)-apply(d01_Ho_B,2,sum)) diffSPO_d0001both_Hs_A <- c(apply(d0001_Hs_A,2,sum)-apply(d001_Hs_A,2,sum),apply(d0001_Hs_A,2,sum)-apply(d01_Hs_A,2,sum)) diffSPO_d0001both_Hs_B <- c(apply(d0001_Hs_B,2,sum)-apply(d001_Hs_B,2,sum),apply(d0001_Hs_B,2,sum)-apply(d01_Hs_B,2,sum)) diffMPO_d0001both_Ar_A <- c(apply(d0001_Ar_A,2,median)-apply(d001_Ar_A,2,median),apply(d0001_Ar_A,2,median)-apply(d01_Ar_A,2,median)) diffMPO_d0001both_Ar_B <- c(apply(d0001_Ar_B,2,median)-apply(d001_Ar_B,2,median),apply(d0001_Ar_B,2,median)-apply(d01_Ar_B,2,median)) diffMPO_d0001both_Ho_A <- c(apply(d0001_Ho_A,2,median)-apply(d001_Ho_A,2,median),apply(d0001_Ho_A,2,median)-apply(d01_Ho_A,2,median)) diffMPO_d0001both_Ho_B <- c(apply(d0001_Ho_B,2,median)-apply(d001_Ho_B,2,median),apply(d0001_Ho_B,2,median)-apply(d01_Ho_B,2,median)) diffMPO_d0001both_Hs_A <- c(apply(d0001_Hs_A,2,median)-apply(d001_Hs_A,2,median),apply(d0001_Hs_A,2,median)-apply(d01_Hs_A,2,median)) diffMPO_d0001both_Hs_B <- c(apply(d0001_Hs_B,2,median)-apply(d001_Hs_B,2,median),apply(d0001_Hs_B,2,median)-apply(d01_Hs_B,2,median)) list_d0001both <- c(diffMPO_d0001both_Ar_A,diffMPO_d0001both_Ar_B, diffMPO_d0001both_Ho_A,diffMPO_d0001both_Ho_B, diffMPO_d0001both_Hs_A,diffMPO_d0001both_Hs_B, diffSPO_d0001both_Ar_A,diffSPO_d0001both_Ar_B, diffSPO_d0001both_Ho_A,diffSPO_d0001both_Ho_B, diffSPO_d0001both_Hs_A,diffSPO_d0001both_Hs_B) total_comparison_d0001both <- length(list_d0001both) improved_fit_d0001both <- sum(list_d0001both<0) # Model performance d improved_fit_d0001both/total_comparison_d0001both ##** FIG S8 #### ##** Perpendicular offset histogram comparison for d if(pdf){ pdf(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS8.pdf"), width=8, height=6) }else{ png(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS8.png"), width=8, height=6, units="in", res=300) } op <- par(mfrow = c(3,3), oma = c(5,5,2,0) + 0.1, mar = c(0,0,2,1) + 0.1) hist(d0001_Ar_A, col=col_Gfos_A, cex.axis=2, xaxt="n", main="", ylim=c(0,70), xlim = c(0,maxhist_Ar), breaks=seq(0,maxhist_Ar,0.25)) hist(d0001_Ar_B, col=col_Gfos_B, add=T, breaks=seq(0,maxhist_Ar,0.25)) axis(1, labels=F, cex.axis=2) mtext(side = 3, text = lab_Ar_short, line = 1, adj=0.5, cex = 1.5) abline(v=median(d0001_Ar_A, na.rm=T), col=white_transparent, lwd=2) abline(v=median(d0001_Ar_A, na.rm=T), col=col_Gfos_A, lty=2, lwd=2) abline(v=median(d0001_Ar_B, na.rm=T), col=white_transparent, lwd=2) abline(v=median(d0001_Ar_B, na.rm=T), col=col_Gfos_B, lty=2, lwd=2) textbox(4,65, txt = lab[6], txt.adj=0.5, txt.cex = 2, frm.brd = NA, frm.col = white_transparent) legend(2,50, c(expression("Median" ~ italic(G. ~ fossarum) ~ "type A"),expression("Median" ~ italic(G. ~ fossarum) ~ "type B")), lty=2, col=c(col_Gfos_A,col_Gfos_B), bty="n", cex=0.85, lwd=2) hist(d0001_Ho_A, col=col_Gfos_A, yaxt="n", xaxt="n", main="", ylim=c(0,70), xlim = c(0,maxhist_Ho), breaks=seq(0,maxhist_Ho,0.025)) hist(d0001_Ho_B, col=col_Gfos_B, add=T, breaks=seq(0,maxhist_Ho,0.025)) axis(1, labels=F, cex.axis=2) axis(2, labels=F, cex.axis=2) mtext(side = 3, text = lab_Ho_short, line = 1, adj=0.5, cex = 1.5) abline(v=median(d0001_Ho_A, na.rm=T), col=white_transparent, lwd=2) abline(v=median(d0001_Ho_A, na.rm=T), col=col_Gfos_A, lty=2, lwd=2) abline(v=median(d0001_Ho_B, na.rm=T), col=white_transparent, lwd=2) abline(v=median(d0001_Ho_B, na.rm=T), col=col_Gfos_B, lty=2, lwd=2) textbox(0.3,65, txt = lab[6], txt.adj=0.5, txt.cex = 2, frm.brd = NA, frm.col = white_transparent) hist(d0001_Hs_A, col=col_Gfos_A, yaxt="n", xaxt="n", main="", ylim=c(0,70), xlim = c(0,maxhist_Hs), breaks=seq(0,maxhist_Hs,0.025)) hist(d0001_Hs_B, col=col_Gfos_B, add=T, breaks=seq(0,maxhist_Hs,0.025)) axis(1, labels=F, cex.axis=2) axis(2, labels=F, cex.axis=2) mtext(side = 3, text = lab_Hs_short, line = 1, adj=0.5, cex = 1.5) abline(v=median(d0001_Hs_A, na.rm=T), col=white_transparent, lwd=2) abline(v=median(d0001_Hs_A, na.rm=T), col=col_Gfos_A, lty=2, lwd=2) abline(v=median(d0001_Hs_B, na.rm=T), col=white_transparent, lwd=2) abline(v=median(d0001_Hs_B, na.rm=T), col=col_Gfos_B, lty=2, lwd=2) textbox(0.3,65, txt = lab[6], txt.adj=0.5, txt.cex = 2, frm.brd = NA, frm.col = white_transparent) hist(d001_Ar_A, col=col_Gfos_A, cex.axis=2, xaxt="n", main="", ylim=c(0,70), xlim = c(0,maxhist_Ar), breaks=seq(0,maxhist_Ar,0.25)) hist(d001_Ar_B, col=col_Gfos_B, add=T, breaks=seq(0,maxhist_Ar,0.25)) axis(1, labels=F, cex.axis=2) abline(v=median(d001_Ar_A, na.rm=T), col=white_transparent, lwd=2) abline(v=median(d001_Ar_A, na.rm=T), col=col_Gfos_A, lty=2, lwd=2) abline(v=median(d001_Ar_B, na.rm=T), col=white_transparent, lwd=2) abline(v=median(d001_Ar_B, na.rm=T), col=col_Gfos_B, lty=2, lwd=2) textbox(4,65, txt = lab[7], txt.adj=0.5, txt.cex = 2, frm.brd = NA, frm.col = white_transparent) hist(d001_Ho_A, col=col_Gfos_A, yaxt="n", xaxt="n", main="", ylim=c(0,70), xlim = c(0,maxhist_Ho), breaks=seq(0,maxhist_Ho,0.025)) hist(d001_Ho_B, col=col_Gfos_B, add=T, breaks=seq(0,maxhist_Ho,0.025)) axis(1, labels=F, cex.axis=2) axis(2, labels=F, cex.axis=2) abline(v=median(d001_Ho_A, na.rm=T), col=white_transparent, lwd=2) abline(v=median(d001_Ho_A, na.rm=T), col=col_Gfos_A, lty=2, lwd=2) abline(v=median(d001_Ho_B, na.rm=T), col=white_transparent, lwd=2) abline(v=median(d001_Ho_B, na.rm=T), col=col_Gfos_B, lty=2, lwd=2) textbox(0.3,65, txt = lab[7], txt.adj=0.5, txt.cex = 2, frm.brd = NA, frm.col = white_transparent) hist(d001_Hs_A, col=col_Gfos_A, yaxt="n", xaxt="n", main="", ylim=c(0,70), xlim = c(0,maxhist_Hs), breaks=seq(0,maxhist_Hs,0.025)) hist(d001_Hs_B, col=col_Gfos_B, add=T, breaks=seq(0,maxhist_Hs,0.025)) axis(1, labels=F, cex.axis=2) axis(2, labels=F, cex.axis=2) abline(v=median(d001_Hs_A, na.rm=T), col=white_transparent, lwd=2) abline(v=median(d001_Hs_A, na.rm=T), col=col_Gfos_A, lty=2, lwd=2) abline(v=median(d001_Hs_B, na.rm=T), col=white_transparent, lwd=2) abline(v=median(d001_Hs_B, na.rm=T), col=col_Gfos_B, lty=2, lwd=2) textbox(0.3,65, txt = lab[7], txt.adj=0.5, txt.cex = 2, frm.brd = NA, frm.col = white_transparent) hist(d01_Ar_A, col=col_Gfos_A, cex.axis=2, xaxt="n", main="", ylim=c(0,70), xlim = c(0,maxhist_Ar), breaks=seq(0,maxhist_Ar,0.25)) hist(d01_Ar_B, col=col_Gfos_B, add=T, breaks=seq(0,maxhist_Ar,0.25)) axis(1, labels=T, cex.axis=2, padj=0.5) abline(v=median(d01_Ar_A, na.rm=T), col=white_transparent, lwd=2) abline(v=median(d01_Ar_A, na.rm=T), col=col_Gfos_A, lty=2, lwd=2) abline(v=median(d01_Ar_B, na.rm=T), col=white_transparent, lwd=2) abline(v=median(d01_Ar_B, na.rm=T), col=col_Gfos_B, lty=2, lwd=2) textbox(4,65, txt = lab[8], txt.adj=0.5, txt.cex = 2, frm.brd = NA, frm.col = white_transparent) hist(d01_Ho_A, col=col_Gfos_A, yaxt="n", xaxt="n", main="", ylim=c(0,70), xlim = c(0,maxhist_Ho), breaks=seq(0,maxhist_Ho,0.025)) hist(d01_Ho_B, col=col_Gfos_B, add=T, breaks=seq(0,maxhist_Ho,0.025)) axis(1, labels=T, cex.axis=2, padj=0.5) axis(2, labels=F, cex.axis=2) abline(v=median(d01_Ho_A, na.rm=T), col=white_transparent, lwd=2) abline(v=median(d01_Ho_A, na.rm=T), col=col_Gfos_A, lty=2, lwd=2) abline(v=median(d01_Ho_B, na.rm=T), col=white_transparent, lwd=2) abline(v=median(d01_Ho_B, na.rm=T), col=col_Gfos_B, lty=2, lwd=2) textbox(0.3,65, txt = lab[8], txt.adj=0.5, txt.cex = 2, frm.brd = NA, frm.col = white_transparent) hist(d01_Hs_A, col=col_Gfos_A, yaxt="n", xaxt="n", main="", ylim=c(0,70), xlim = c(0,maxhist_Hs), breaks=seq(0,maxhist_Hs,0.025)) hist(d01_Hs_B, col=col_Gfos_B, add=T, breaks=seq(0,maxhist_Hs,0.025)) axis(1, labels=T, cex.axis=2, padj=0.5) axis(2, labels=F, cex.axis=2) abline(v=median(d01_Hs_A, na.rm=T), col=white_transparent, lwd=2) abline(v=median(d01_Hs_A, na.rm=T), col=col_Gfos_A, lty=2, lwd=2) abline(v=median(d01_Hs_B, na.rm=T), col=white_transparent, lwd=2) abline(v=median(d01_Hs_B, na.rm=T), col=col_Gfos_B, lty=2, lwd=2) textbox(0.3,65, txt = lab[8], txt.adj=0.5, txt.cex = 2, frm.brd = NA, frm.col = white_transparent) title(xlab = "Perpendicular offset", ylab = "Frequency", outer = TRUE, line = 3.5, cex.lab=2) dev.off() ##** K: Model performance #### k0_Ar_A <- hist_Ar_A[,c(1,3,5,7,9,11,13,15,17)] k1_Ar_A <- hist_Ar_A[,c(2,4,6,8,10,12,14,16,18)] k0_Ar_B <- hist_Ar_B[,c(1,3,5,7,9,11,13,15,17)] k1_Ar_B <- hist_Ar_B[,c(2,4,6,8,10,12,14,16,18)] k0_Ho_A <- hist_Ho_A[,c(1,3,5,7,9,11,13,15,17)] k1_Ho_A <- hist_Ho_A[,c(2,4,6,8,10,12,14,16,18)] k0_Ho_B <- hist_Ho_B[,c(1,3,5,7,9,11,13,15,17)] k1_Ho_B <- hist_Ho_B[,c(2,4,6,8,10,12,14,16,18)] k0_Hs_A <- hist_Hs_A[,c(1,3,5,7,9,11,13,15,17)] k1_Hs_A <- hist_Hs_A[,c(2,4,6,8,10,12,14,16,18)] k0_Hs_B <- hist_Hs_B[,c(1,3,5,7,9,11,13,15,17)] k1_Hs_B <- hist_Hs_B[,c(2,4,6,8,10,12,14,16,18)] diffSPO_k0k1_Ar_A <- apply(k0_Ar_A,2,sum)-apply(k1_Ar_A,2,sum) diffSPO_k0k1_Ar_B <- apply(k0_Ar_B,2,sum)-apply(k1_Ar_B,2,sum) diffSPO_k0k1_Ho_A <- apply(k0_Ho_A,2,sum)-apply(k1_Ho_A,2,sum) diffSPO_k0k1_Ho_B <- apply(k0_Ho_B,2,sum)-apply(k1_Ho_B,2,sum) diffSPO_k0k1_Hs_A <- apply(k0_Hs_A,2,sum)-apply(k1_Hs_A,2,sum) diffSPO_k0k1_Hs_B <- apply(k0_Hs_B,2,sum)-apply(k1_Hs_B,2,sum) diffMPO_k0k1_Ar_A <- apply(k0_Ar_A,2,median)-apply(k1_Ar_A,2,median) diffMPO_k0k1_Ar_B <- apply(k0_Ar_B,2,median)-apply(k1_Ar_B,2,median) diffMPO_k0k1_Ho_A <- apply(k0_Ho_A,2,median)-apply(k1_Ho_A,2,median) diffMPO_k0k1_Ho_B <- apply(k0_Ho_B,2,median)-apply(k1_Ho_B,2,median) diffMPO_k0k1_Hs_A <- apply(k0_Hs_A,2,median)-apply(k1_Hs_A,2,median) diffMPO_k0k1_Hs_B <- apply(k0_Hs_B,2,median)-apply(k1_Hs_B,2,median) list_k0k1 <- c(diffMPO_k0k1_Ar_A,diffMPO_k0k1_Ar_B, diffMPO_k0k1_Ho_A,diffMPO_k0k1_Ho_B, diffMPO_k0k1_Hs_A,diffMPO_k0k1_Hs_B, diffSPO_k0k1_Ar_A,diffSPO_k0k1_Ar_B, diffSPO_k0k1_Ho_A,diffSPO_k0k1_Ho_B, diffSPO_k0k1_Hs_A,diffSPO_k0k1_Hs_B) total_comparison_k0k1 <- length(list_k0k1) improved_fit_k0k1 <- sum(list_k0k1<0) # Model performance K improved_fit_k0k1/total_comparison_k0k1 ##** FIG S9 #### ##** Perpendicular offset histogram comparison for K scal.fact <- 0.75 if(pdf){ pdf(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS9.pdf"), width=8, height=6*scal.fact) }else{ png(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS9.png"), width=8, height=6*scal.fact, units="in", res=300) } op <- par(mfrow = c(3*scal.fact,3), oma = c(5*scal.fact,5*scal.fact,2*scal.fact,0) + 0.1, mar = c(0,0,2*scal.fact,1*scal.fact) + 0.1) hist(k0_Ar_A, col=col_Gfos_A, tcl=-0.5*scal.fact, cex.axis=2*scal.fact, xaxt="n", yaxt="n", main="", ylim=c(0,90), xlim = c(0,maxhist_Ar), breaks=seq(0,maxhist_Ar,0.25)) hist(k0_Ar_B, col=col_Gfos_B, add=T, breaks=seq(0,maxhist_Ar,0.25)) axis(1, labels=F, cex.axis=2*scal.fact, tcl=-0.5*scal.fact) axis(2, labels=T, cex.axis=2*scal.fact, tcl=-0.5*scal.fact, padj=0.5) mtext(side = 3, text = lab_Ar_short, line = 1, adj=0.5, cex = 1.5) abline(v=median(k0_Ar_A, na.rm=T), col=white_transparent, lwd=2) abline(v=median(k0_Ar_A, na.rm=T), col=col_Gfos_A, lty=2, lwd=2) abline(v=median(k0_Ar_B, na.rm=T), col=white_transparent, lwd=2) abline(v=median(k0_Ar_B, na.rm=T), col=col_Gfos_B, lty=2, lwd=2) textbox(4,85, txt = lab[1], txt.adj=0.5, txt.cex = 2*scal.fact, frm.brd = NA, frm.col = white_transparent) legend(2,70, c(expression("Median" ~ italic(G. ~ fossarum) ~ "type A"),expression("Median" ~ italic(G. ~ fossarum) ~ "type B")), lty=2, col=c(col_Gfos_A,col_Gfos_B), bty="n", cex=0.85, lwd=2) hist(k0_Ho_A, col=col_Gfos_A, tcl=-0.5*scal.fact, cex.axis=2*scal.fact, xaxt="n", yaxt="n", main="", ylim=c(0,90), xlim = c(0,maxhist_Ho), breaks=seq(0,maxhist_Ho,0.025)) hist(k0_Ho_B, col=col_Gfos_B, add=T, breaks=seq(0,maxhist_Ho,0.025)) axis(1, labels=F, cex.axis=2*scal.fact, tcl=-0.5*scal.fact) axis(2, labels=F, cex.axis=2*scal.fact, tcl=-0.5*scal.fact) mtext(side = 3, text = lab_Ho_short, line = 1, adj=0.5, cex = 1.5) abline(v=median(k0_Ho_A, na.rm=T), col=white_transparent, lwd=2) abline(v=median(k0_Ho_A, na.rm=T), col=col_Gfos_A, lty=2, lwd=2) abline(v=median(k0_Ho_B, na.rm=T), col=white_transparent, lwd=2) abline(v=median(k0_Ho_B, na.rm=T), col=col_Gfos_B, lty=2, lwd=2) textbox(0.3,85, txt = lab[1], txt.adj=0.5, txt.cex = 2*scal.fact, frm.brd = NA, frm.col = white_transparent) hist(k0_Hs_A, col=col_Gfos_A, tcl=-0.5*scal.fact, cex.axis=2*scal.fact, xaxt="n", yaxt="n", main="", ylim=c(0,90), xlim = c(0,maxhist_Hs), breaks=seq(0,maxhist_Hs,0.025)) hist(k0_Hs_B, col=col_Gfos_B, add=T, breaks=seq(0,maxhist_Hs,0.025)) axis(1, labels=F, cex.axis=2*scal.fact, tcl=-0.5*scal.fact) axis(2, labels=F, cex.axis=2*scal.fact, tcl=-0.5*scal.fact) mtext(side = 3, text = lab_Hs_short, line = 1, adj=0.5, cex = 1.5) abline(v=median(k0_Hs_A, na.rm=T), col=white_transparent, lwd=2) abline(v=median(k0_Hs_A, na.rm=T), col=col_Gfos_A, lty=2, lwd=2) abline(v=median(k0_Hs_B, na.rm=T), col=white_transparent, lwd=2) abline(v=median(k0_Hs_B, na.rm=T), col=col_Gfos_B, lty=2, lwd=2) textbox(0.3,85, txt = lab[1], txt.adj=0.5, txt.cex = 2*scal.fact, frm.brd = NA, frm.col = white_transparent) hist(k1_Ar_A, col=col_Gfos_A, tcl=-0.5*scal.fact, cex.axis=2*scal.fact, xaxt="n", yaxt="n", main="", ylim=c(0,90), xlim = c(0,maxhist_Ar), breaks=seq(0,maxhist_Ar,0.25)) hist(k1_Ar_B, col=col_Gfos_B, add=T, breaks=seq(0,maxhist_Ar,0.25)) axis(1, labels=T, cex.axis=2*scal.fact, tcl=-0.5*scal.fact, padj=0) axis(2, labels=T, cex.axis=2*scal.fact, tcl=-0.5*scal.fact, padj=0.5) abline(v=median(k1_Ar_A, na.rm=T), col=white_transparent, lwd=2) abline(v=median(k1_Ar_A, na.rm=T), col=col_Gfos_A, lty=2, lwd=2) abline(v=median(k1_Ar_B, na.rm=T), col=white_transparent, lwd=2) abline(v=median(k1_Ar_B, na.rm=T), col=col_Gfos_B, lty=2, lwd=2) textbox(4,85, txt = lab[2], txt.adj=0.5, txt.cex = 2*scal.fact, frm.brd = NA, frm.col = white_transparent) hist(k1_Ho_A, col=col_Gfos_A, tcl=-0.5*scal.fact, cex.axis=2*scal.fact, xaxt="n", yaxt="n", main="", ylim=c(0,90), xlim = c(0,maxhist_Ho), breaks=seq(0,maxhist_Ho,0.025)) hist(k1_Ho_B, col=col_Gfos_B, add=T, breaks=seq(0,maxhist_Ho,0.025)) axis(1, labels=T, cex.axis=2*scal.fact, tcl=-0.5*scal.fact, padj=0) axis(2, labels=F, cex.axis=2*scal.fact, tcl=-0.5*scal.fact) abline(v=median(k1_Ho_A, na.rm=T), col=white_transparent, lwd=2) abline(v=median(k1_Ho_A, na.rm=T), col=col_Gfos_A, lty=2, lwd=2) abline(v=median(k1_Ho_B, na.rm=T), col=white_transparent, lwd=2) abline(v=median(k1_Ho_B, na.rm=T), col=col_Gfos_B, lty=2, lwd=2) textbox(0.3,85, txt = lab[2], txt.adj=0.5, txt.cex = 2*scal.fact, frm.brd = NA, frm.col = white_transparent) hist(k1_Hs_A, col=col_Gfos_A, tcl=-0.5*scal.fact, cex.axis=2*scal.fact, xaxt="n", yaxt="n", main="", ylim=c(0,90), xlim = c(0,maxhist_Hs), breaks=seq(0,maxhist_Hs,0.025)) hist(k1_Hs_B, col=col_Gfos_B, add=T, breaks=seq(0,maxhist_Hs,0.025)) axis(1, labels=T, cex.axis=2*scal.fact, tcl=-0.5*scal.fact, padj=0) axis(2, labels=F, cex.axis=2*scal.fact, tcl=-0.5*scal.fact) abline(v=median(k1_Hs_A, na.rm=T), col=white_transparent, lwd=2) abline(v=median(k1_Hs_A, na.rm=T), col=col_Gfos_A, lty=2, lwd=2) abline(v=median(k1_Hs_B, na.rm=T), col=white_transparent, lwd=2) abline(v=median(k1_Hs_B, na.rm=T), col=col_Gfos_B, lty=2, lwd=2) textbox(0.3,85, txt = lab[2], txt.adj=0.5, txt.cex = 2*scal.fact, frm.brd = NA, frm.col = white_transparent) title(xlab = "Perpendicular offset", ylab = "Frequency", outer = TRUE, line = 3.5*scal.fact, cex.lab=2*scal.fact) dev.off() ##** FIG S10 #### ##** Orthogonal distance to 1:1 line mp_dim <- multipanel.dimensions(main.col=pla$a,main.row=pla$b,pla$x,pla$y,sub1=pla$sub1,sub2=pla$sub2,main=pla$main) if(pdf){ pdf(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS10.pdf"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1])) }else{ png(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS10.png"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1]), units="in", res=600) } nf <- multipanel.layout(main.col=pla$a,main.row=pla$b,pla$x,pla$y,sub1=pla$sub1,sub2=pla$sub2,main=pla$main) for (i in 2:ncol(Ar_modelled)){ Ar_Mod_A <- Ar_modelled[,i][rownames(Ar_modelled)%in%modsite_GfosA] Ar_Mod_B <- Ar_modelled[,i][rownames(Ar_modelled)%in%modsite_GfosB] par(mar=c(0,0,0,0)) plot(Ar_Mod_A~meanAr_A_red_updist, type="n", xlim=c(min(Ar_modelled[,-1], na.rm=T),max(Ar_modelled[,-1], na.rm=T)), ylim=c(min(Ar_modelled[,-1], na.rm=T),max(Ar_modelled[,-1], na.rm=T)), xaxt=ifelse(i%in%c(6,7,12,13,18,19), "s", "n"), yaxt=ifelse(i%in%c(3,5,7), "s", "n"), asp=1) # create empty plot if (i %in% c(14,16,18)){ axis(4, labels=F) } if (i %in% c(2,4,6)){ axis(2, labels=F) } if (i %in% c(2,3)){ mtext(lab[6], side=3, line=0.8, cex=1.4) } if (i %in% c(8,9)){ mtext(lab[7], side=3, line=0.8, cex=1.4) } if (i %in% c(14,15)){ mtext(lab[8], side=3, line=0.8, cex=1.4) } if (i %in% c(4)){ mtext(expression(bold("Model data: ")*" Mean allelic richness"), side=2, line=1, cex=1) } if (i %in% c(12,13)){ mtext(expression(bold("Empirical data:")*" Mean allelic richness"), side=1, line=3, cex=1) } abline(0,1, lwd=1, lty=2) # add 1:1 line points(Ar_Mod_A~meanAr_A_red_updist,col=col_Gfos_A, pch=16) points(Ar_Mod_B~meanAr_B_updist,col=col_Gfos_B, pch=16) for (j in 1:length(meanAr_A_red_updist)){ point <- cbind(meanAr_A_red_updist,Ar_Mod_A)[j,] seg <- unlist(perp.segment.coord(point[1],point[2])) segments(seg[1],seg[2],seg[3],seg[4], col=col_Gfos_A, lty=2, lwd=0.5) } for (k in 1:length(meanAr_B_updist)){ point <- cbind(meanAr_B_updist,Ar_Mod_B)[k,] seg <- unlist(perp.segment.coord(point[1],point[2])) segments(seg[1],seg[2],seg[3],seg[4], col=col_Gfos_B, lty=2, lwd=0.5) } textbox(max(Ar_modelled[,-1], na.rm=T),min(Ar_modelled[,-1], na.rm=T)+2,paste0(measure1_short, " = ",formatC(round(sum_orthdist_Ar_A[[i-1]],sum_digits),digits=sum_digits, format="f")), txt.cex=sum_cex, txt.adj=1, txt.col=col_Gfos_A, frm.col=white_transparent, frm.brd = NA, frm.siz = 0.2) textbox(max(Ar_modelled[,-1], na.rm=T),min(Ar_modelled[,-1], na.rm=T)+0.5,paste0(measure1_short, " = ",formatC(round(sum_orthdist_Ar_B[[i-1]],sum_digits),digits=sum_digits, format="f")), txt.cex=sum_cex, txt.adj=1, txt.col=col_Gfos_B, frm.col=white_transparent, frm.brd = NA, frm.siz = 0.2) } for (i in 1:length(lab)){ plot.new() if (i %in% c(1:2)){ text(0.5,1,lab[i],adj=c(0.5,1), cex=lab_cex[i]) } if (i %in% c(3:5)){ text(0,0.5,lab[i],adj=c(0,0.5), cex=lab_cex[i]) } } plot.new() legend("topleft",c(label_A,label_B),pch = 16, col = c(col_Gfos_A,col_Gfos_B), bty="n", cex=2) text(0,0,measure1, adj=0) dev.off() ##** TABLE S11 #### SPOrank_Ar_A <- labs_comb[order(apply(hist_Ar_A,2,sum))] SPOrank_Ar_B <- labs_comb[order(apply(hist_Ar_B,2,sum))] SPOrank_spo_Ar_A <- apply(hist_Ar_A,2,sum)[order(apply(hist_Ar_A,2,sum))] SPOrank_d_Ar_A <- labs_short[,3][order(apply(hist_Ar_A,2,sum))] SPOrank_W_Ar_A <- labs_short[,2][order(apply(hist_Ar_A,2,sum))] SPOrank_K_Ar_A <- labs_short[,1][order(apply(hist_Ar_A,2,sum))] SPOrank_spo_Ar_B <- apply(hist_Ar_B,2,sum)[order(apply(hist_Ar_B,2,sum))] SPOrank_d_Ar_B <- labs_short[,3][order(apply(hist_Ar_B,2,sum))] SPOrank_W_Ar_B <- labs_short[,2][order(apply(hist_Ar_B,2,sum))] SPOrank_K_Ar_B <- labs_short[,1][order(apply(hist_Ar_B,2,sum))] SPOrank_Ar <- data.frame("Ar_A_SPO"=SPOrank_spo_Ar_A,"Ar_A_d"=SPOrank_d_Ar_A,"Ar_A_W"=SPOrank_W_Ar_A,"Ar_A_K"=SPOrank_K_Ar_A, "Ar_B_SPO"=SPOrank_spo_Ar_B,"Ar_B_d"=SPOrank_d_Ar_B,"Ar_B_W"=SPOrank_W_Ar_B,"Ar_B_K"=SPOrank_K_Ar_B) write.csv2(SPOrank_Ar, paste0(WD,"/Analysis_",output,"/SuppFigs/Table_S11.csv")) ##** FIG S12 #### ###** Histogram of orthogonal distance to 1:1 line main_5=F mp_dim <- multipanel.dimensions(main.col=pla$a,main.row=pla$b,pla$x,pla$y,sub1=pla$sub1,sub2=pla$sub2,main=main_5, h.main=0.5) if(pdf){ pdf(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS12.pdf"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1])) }else{ png(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS12.png"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1]), units="in", res=600) } nf <- multipanel.layout(main.col=pla$a,main.row=pla$b,pla$x,pla$y,sub1=pla$sub1,sub2=pla$sub2,main=main_5, h.main=0.5) for (i in 1:ncol(hist_Ho_A)){ par(mar=c(0,0,0,0)) yclip <- 28 ylim <- 35 hist(hist_Ho_A[,i], breaks=seq(0,ceiling(max(hist_Ho_A,hist_Ho_B, na.rm=T)/0.1)*0.1,0.05), xlim=c(0,ceiling(max(hist_Ho_A,hist_Ho_B, na.rm=T)/0.1)*0.1), ylim=c(0,ylim), xaxt="n", yaxt="n", col=col_Gfos_A, main="") clip(0,round(max(hist_Ho_A,hist_Ho_B, na.rm=T),1), 0, yclip) abline(v=median(hist_Ho_A[,i], na.rm=T),col=col_Gfos_A) abline(v=median(hist_Ho_B[,i], na.rm=T),col=col_Gfos_B) clip(0,round(max(hist_Ho_A,hist_Ho_B, na.rm=T),1), 0, ylim) textbox(round(max(hist_Ho_A,hist_Ho_B, na.rm=T),1),30,paste0(measure2_short," = ",formatC(round(median(hist_Ho_A[,i], na.rm=T),median_digits),digits=median_digits, format="f")), txt.cex=median_cex, txt.col=col_Gfos_A, txt.adj=1, frm.col="white", frm.brd = "white", frm.siz = 0.2) textbox(round(max(hist_Ho_A,hist_Ho_B, na.rm=T),1),25,paste0(measure2_short," = ",formatC(round(median(hist_Ho_B[,i], na.rm=T),median_digits),digits=median_digits, format="f")), txt.cex=median_cex, txt.col=col_Gfos_B, txt.adj=1, frm.col="white", frm.brd = "white", frm.siz = 0.2) hist(hist_Ho_A[,i], breaks=seq(0,ceiling(max(hist_Ho_A,hist_Ho_B, na.rm=T)/0.1)*0.1,0.05), xlim=c(0,ceiling(max(hist_Ho_A,hist_Ho_B, na.rm=T)/0.1)*0.1), ylim=c(0,ylim), xaxt="n", yaxt="n", col=col_Gfos_A, main="", add=T) hist(hist_Ho_B[,i], breaks=seq(0,ceiling(max(hist_Ho_A,hist_Ho_B, na.rm=T)/0.1)*0.1,0.05), xlim=c(0,ceiling(max(hist_Ho_A,hist_Ho_B, na.rm=T)/0.1)*0.1), ylim=c(0,ylim), xaxt="n", yaxt="n", col=col_Gfos_B, main="", add=T) if (i %in% c(1,2)){ mtext(lab[6], side=3, line=0.8, cex=1.3) } if (i %in% c(7,8)){ mtext(lab[7], side=3, line=0.8, cex=1.3) } if (i %in% c(13,14)){ mtext(lab[8], side=3, line=0.8, cex=1.3) } if (i %in% c(3)){ mtext("Frequencies [counts]", side=2, line=1, cex=1) } if (i %in% c(11,12)){ mtext(measure3, side=1, line=3, cex=1) } if (i %in% c(5:6,11:12,17:18)){ axis(1, at=seq(0,round(max(hist_Ho_A,hist_Ho_B, na.rm=T),1),0.1), labels=c("0.0","","0.2","","0.4","","")) } if (i %in% c(1,3,5)){ axis(2, at=seq(0,yclip,5), labels=F) } if (i %in% c(2,4,6)){ axis(2, at=seq(0,yclip,5)) } if (i %in% c(13:18)){ axis(4, at=seq(0,yclip,5), labels=F) } } if(main_5){ plot.new() text(0.5,0.7,paste0(lab_Ho,": Distribution of ",measure3a),adj=c(0.5,0.5),cex=3) } for (i in 1:length(lab)){ plot.new() if (i %in% c(1:2)){ text(0.5,1,lab[i],adj=c(0.5,1), cex=lab_cex[i]) } if (i %in% c(3:5)){ text(0,0.5,lab[i],adj=c(0,0.5), cex=lab_cex[i]) } } par(mar=c(0,6,16,0.5), pty="s") i <- ncol(Ar_modelled) Ar_Mod_A <- Ar_modelled[,i][rownames(Ar_modelled)%in%modsite_GfosA] Ar_Mod_B <- Ar_modelled[,i][rownames(Ar_modelled)%in%modsite_GfosB] plot(Ar_Mod_A~meanAr_A_red_updist, type="n", xlim=c(min(Ar_modelled[,-1], na.rm=T),max(Ar_modelled[,-1], na.rm=T)), ylim=c(min(Ar_modelled[,-1], na.rm=T),max(Ar_modelled[,-1], na.rm=T)), xaxt="n", yaxt="n", xlab="", ylab="", bty="n", asp=1) par(xpd=T) legend(-5,yclip,c(label_A,label_B),pch = 16, col = c(col_Gfos_A,col_Gfos_B), bty="n", cex=2) par(xpd=F) dev.off() ##** FIG S13 #### ##** Orthogonal distance to 1:1 line mp_dim <- multipanel.dimensions(main.col=pla$a,main.row=pla$b,pla$x,pla$y,sub1=pla$sub1,sub2=pla$sub2,main=pla$main) if(pdf){ pdf(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS13.pdf"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1])) }else{ png(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS13.png"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1]), units="in", res=600) } nf <- multipanel.layout(main.col=pla$a,main.row=pla$b,pla$x,pla$y,sub1=pla$sub1,sub2=pla$sub2,main=pla$main) for (i in 2:ncol(Ho_modelled)){ Ho_Mod_A <- Ho_modelled[,i][rownames(Ho_modelled)%in%modsite_GfosA] Ho_Mod_B <- Ho_modelled[,i][rownames(Ho_modelled)%in%modsite_GfosB] par(mar=c(0,0,0,0)) plot(Ho_Mod_A~meanHo_A_red_updist, type="n", xlim=c(min(Ho_modelled[,-1], na.rm=T),max(Ho_modelled[,-1], na.rm=T)), ylim=c(min(Ho_modelled[,-1], na.rm=T),max(Ho_modelled[,-1], na.rm=T)), xaxt=ifelse(i%in%c(6,7,12,13,18,19), "s", "n"), yaxt=ifelse(i%in%c(3,5,7), "s", "n"), asp=1) # create empty plot if (i %in% c(14,16,18)){ axis(4, labels=F) } if (i %in% c(2,4,6)){ axis(2, labels=F) } if (i %in% c(2,3)){ mtext(lab[6], side=3, line=0.8, cex=1.3) } if (i %in% c(8,9)){ mtext(lab[7], side=3, line=0.8, cex=1.3) } if (i %in% c(14,15)){ mtext(lab[8], side=3, line=0.8, cex=1.3) } if (i %in% c(4)){ mtext(expression(bold("Model data:")*" Mean observed heterozygosity"), side=2, line=1, cex=1) } if (i %in% c(12,13)){ mtext(expression(bold("Empirical data:")*" Mean observed heterozygosity"), side=1, line=3, cex=1) } abline(0,1, lwd=1, lty=2) # add 1:1 line points(Ho_Mod_A~meanHo_A_red_updist,col=col_Gfos_A, pch=16) points(Ho_Mod_B~meanHo_B_updist,col=col_Gfos_B, pch=16) for (j in 1:length(meanHo_A_red_updist)){ point <- cbind(meanHo_A_red_updist,Ho_Mod_A)[j,] seg <- unlist(perp.segment.coord(point[1],point[2])) segments(seg[1],seg[2],seg[3],seg[4], col=col_Gfos_A, lty=2, lwd=0.5) } for (k in 1:length(meanHo_B_updist)){ point <- cbind(meanHo_B_updist,Ho_Mod_B)[k,] seg <- unlist(perp.segment.coord(point[1],point[2])) segments(seg[1],seg[2],seg[3],seg[4], col=col_Gfos_B, lty=2, lwd=0.5) } textbox(max(Ho_modelled[,-1], na.rm=T),min(Ho_modelled[,-1], na.rm=T)+0.18,paste0(measure1_short," = ",formatC(round(sum_orthdist_Ho_A[[i-1]],sum_digits),digits=sum_digits, format="f")), txt.cex=sum_cex, txt.adj=1, txt.col=col_Gfos_A, frm.col=white_transparent, frm.brd = NA, frm.siz = 0.2) textbox(max(Ho_modelled[,-1], na.rm=T),min(Ho_modelled[,-1], na.rm=T)+0.05,paste0(measure1_short," = ",formatC(round(sum_orthdist_Ho_B[[i-1]],sum_digits),digits=sum_digits, format="f")), txt.cex=sum_cex, txt.adj=1, txt.col=col_Gfos_B, frm.col=white_transparent, frm.brd = NA, frm.siz = 0.2) } for (i in 1:length(lab)){ plot.new() if (i %in% c(1:2)){ text(0.5,1,lab[i],adj=c(0.5,1), cex=lab_cex[i]) } if (i %in% c(3:5)){ text(0,0.5,lab[i],adj=c(0,0.5), cex=lab_cex[i]) } } plot.new() legend("topleft",c(label_A,label_B),pch = 16, col = c(col_Gfos_A,col_Gfos_B), bty="n", cex=2) text(0,0,measure1, adj=0) dev.off() ##** TABLE S14 #### SPOrank_Ho_A <- labs_comb[order(apply(hist_Ho_A,2,sum))] SPOrank_Ho_B <- labs_comb[order(apply(hist_Ho_B,2,sum))] SPOrank_spo_Ho_A <- apply(hist_Ho_A,2,sum)[order(apply(hist_Ho_A,2,sum))] SPOrank_d_Ho_A <- labs_short[,3][order(apply(hist_Ho_A,2,sum))] SPOrank_W_Ho_A <- labs_short[,2][order(apply(hist_Ho_A,2,sum))] SPOrank_K_Ho_A <- labs_short[,1][order(apply(hist_Ho_A,2,sum))] SPOrank_spo_Ho_B <- apply(hist_Ho_B,2,sum)[order(apply(hist_Ho_B,2,sum))] SPOrank_d_Ho_B <- labs_short[,3][order(apply(hist_Ho_B,2,sum))] SPOrank_W_Ho_B <- labs_short[,2][order(apply(hist_Ho_B,2,sum))] SPOrank_K_Ho_B <- labs_short[,1][order(apply(hist_Ho_B,2,sum))] SPOrank_Ho <- data.frame("Ho_A_SPO"=SPOrank_spo_Ho_A,"Ho_A_d"=SPOrank_d_Ho_A,"Ho_A_W"=SPOrank_W_Ho_A,"Ho_A_K"=SPOrank_K_Ho_A, "Ho_B_SPO"=SPOrank_spo_Ho_B,"Ho_B_d"=SPOrank_d_Ho_B,"Ho_B_W"=SPOrank_W_Ho_B,"Ho_B_K"=SPOrank_K_Ho_B) write.csv2(SPOrank_Ho, paste0(WD,"/Analysis_",output,"/SuppFigs/Table_S14.csv")) ##** FIG S15 #### ##** Histogram of orthogonal distance to 1:1 line main_s15=F mp_dim <- multipanel.dimensions(main.col=pla$a,main.row=pla$b,pla$x,pla$y,sub1=pla$sub1,sub2=pla$sub2,main=main_s15, h.main=0.5) if(pdf){ pdf(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS15.pdf"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1])) }else{ png(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS15.png"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1]), units="in", res=600) } nf <- multipanel.layout(main.col=pla$a,main.row=pla$b,pla$x,pla$y,sub1=pla$sub1,sub2=pla$sub2,main=main_s15, h.main=0.5) for (i in 1:ncol(hist_Hs_A)){ par(mar=c(0,0,0,0)) yclip <- 28 ylim <- 35 hist(hist_Hs_A[,i], breaks=seq(0,ceiling_dec(max(hist_Hs_A,hist_Hs_B, na.rm=T),1),0.05), xlim=c(0,ceiling_dec(max(hist_Hs_A,hist_Hs_B, na.rm=T),1)), ylim=c(0,ylim), xaxt="n", yaxt="n", col=col_Gfos_A, main="") clip(0,ceiling_dec(max(hist_Hs_A,hist_Hs_B, na.rm=T),1), 0, yclip) abline(v=median(hist_Hs_A[,i], na.rm=T),col=col_Gfos_A) abline(v=median(hist_Hs_B[,i], na.rm=T),col=col_Gfos_B) clip(0,ceiling_dec(max(hist_Hs_A,hist_Hs_B, na.rm=T),1), 0, ylim) textbox(ceiling_dec(max(hist_Hs_A,hist_Hs_B, na.rm=T),1),30,paste0(measure2_short," = ",formatC(round(median(hist_Hs_A[,i], na.rm=T),median_digits),digits=median_digits, format="f")), txt.cex=median_cex, txt.col=col_Gfos_A, txt.adj=1, frm.col="white", frm.brd = "white", frm.siz = 0.2) textbox(ceiling_dec(max(hist_Hs_A,hist_Hs_B, na.rm=T),1),25,paste0(measure2_short," = ",formatC(round(median(hist_Hs_B[,i], na.rm=T),median_digits),digits=median_digits, format="f")), txt.cex=median_cex, txt.col=col_Gfos_B, txt.adj=1, frm.col="white", frm.brd = "white", frm.siz = 0.2) hist(hist_Hs_A[,i], breaks=seq(0,ceiling_dec(max(hist_Hs_A,hist_Hs_B, na.rm=T),1),0.05), xlim=c(0,ceiling_dec(max(hist_Hs_A,hist_Hs_B, na.rm=T),1)), ylim=c(0,ylim), xaxt="n", yaxt="n", col=col_Gfos_A, main="", add=T) hist(hist_Hs_B[,i], breaks=seq(0,ceiling_dec(max(hist_Hs_A,hist_Hs_B, na.rm=T),1),0.05), xlim=c(0,ceiling_dec(max(hist_Hs_A,hist_Hs_B, na.rm=T),1)), ylim=c(0,ylim), xaxt="n", yaxt="n", col=col_Gfos_B, main="", add=T) if (i %in% c(1,2)){ mtext(lab[6], side=3, line=0.8, cex=1.3) } if (i %in% c(7,8)){ mtext(lab[7], side=3, line=0.8, cex=1.3) } if (i %in% c(13,14)){ mtext(lab[8], side=3, line=0.8, cex=1.3) } if (i %in% c(3)){ mtext("Frequencies [counts]", side=2, line=1, cex=1) } if (i %in% c(11,12)){ mtext(measure3, side=1, line=3, cex=1) } if (i %in% c(5:6,11:12,17:18)){ axis(1, at=seq(0,ceiling_dec(max(hist_Hs_A,hist_Hs_B, na.rm=T),1),0.1), labels=c("0.0","","0.2","","0.4","","")[1:length(seq(0,ceiling_dec(max(hist_Hs_A,hist_Hs_B, na.rm=T),1),0.1))]) } if (i %in% c(1,3,5)){ axis(2, at=seq(0,yclip,5), labels=F) } if (i %in% c(2,4,6)){ axis(2, at=seq(0,yclip,5)) } if (i %in% c(13:18)){ axis(4, at=seq(0,yclip,5), labels=F) } } if(main_s15){ plot.new() text(0.5,0.7,paste0(lab_Hs,": Distribution of ",measure3a),adj=c(0.5,0.5),cex=3) } for (i in 1:length(lab)){ plot.new() if (i %in% c(1:2)){ text(0.5,1,lab[i],adj=c(0.5,1), cex=lab_cex[i]) } if (i %in% c(3:5)){ text(0,0.5,lab[i],adj=c(0,0.5), cex=lab_cex[i]) } } par(mar=c(0,6,16,0.5), pty="s") i <- ncol(Ar_modelled) Ar_Mod_A <- Ar_modelled[,i][rownames(Ar_modelled)%in%modsite_GfosA] Ar_Mod_B <- Ar_modelled[,i][rownames(Ar_modelled)%in%modsite_GfosB] plot(Ar_Mod_A~meanAr_A_red_updist, type="n", xlim=c(min(Ar_modelled[,-1], na.rm=T),max(Ar_modelled[,-1], na.rm=T)), ylim=c(min(Ar_modelled[,-1], na.rm=T),max(Ar_modelled[,-1], na.rm=T)), xaxt="n", yaxt="n", xlab="", ylab="", bty="n", asp=1) par(xpd=T) legend(-5,yclip,c(label_A,label_B),pch = 16, col = c(col_Gfos_A,col_Gfos_B), bty="n", cex=2) par(xpd=F) dev.off() ##** FIG S16 #### ##** Orthogonal distance to 1:1 line mp_dim <- multipanel.dimensions(main.col=pla$a,main.row=pla$b,pla$x,pla$y,sub1=pla$sub1,sub2=pla$sub2,main=pla$main) if(pdf){ pdf(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS16.pdf"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1])) }else{ png(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS16.png"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1]), units="in", res=600) } nf <- multipanel.layout(main.col=pla$a,main.row=pla$b,pla$x,pla$y,sub1=pla$sub1,sub2=pla$sub2,main=pla$main) for (i in 2:ncol(Hs_modelled)){ Hs_Mod_A <- Hs_modelled[,i][rownames(Hs_modelled)%in%modsite_GfosA] Hs_Mod_B <- Hs_modelled[,i][rownames(Hs_modelled)%in%modsite_GfosB] par(mar=c(0,0,0,0)) plot(Hs_Mod_A~meanHs_A_red_updist, type="n", xlim=c(min(Hs_modelled[,-1],meanHs_A_red_updist,meanHs_B_updist, na.rm=T),max(Hs_modelled[,-1],meanHs_A_red_updist,meanHs_B_updist, na.rm=T)), ylim=c(min(Hs_modelled[,-1],meanHs_A_red_updist,meanHs_B_updist, na.rm=T),max(Hs_modelled[,-1],meanHs_A_red_updist,meanHs_B_updist, na.rm=T)), xaxt=ifelse(i%in%c(6,7,12,13,18,19), "s", "n"), yaxt=ifelse(i%in%c(3,5,7), "s", "n"), asp=1) # create empty plot if (i %in% c(14,16,18)){ axis(4, labels=F) } if (i %in% c(2,4,6)){ axis(2, labels=F) } if (i %in% c(2,3)){ mtext(lab[6], side=3, line=0.8, cex=1.3) } if (i %in% c(8,9)){ mtext(lab[7], side=3, line=0.8, cex=1.3) } if (i %in% c(14,15)){ mtext(lab[8], side=3, line=0.8, cex=1.3) } if (i %in% c(4)){ mtext(expression(bold("Model data:")*" Expected heterozygosity"), side=2, line=1, cex=1) } if (i %in% c(12,13)){ mtext(expression(bold("Empirical data:")*" Expected heterozygosity"), side=1, line=3, cex=1) } abline(0,1, lwd=1, lty=2) # add 1:1 line points(Hs_Mod_A~meanHs_A_red_updist,col=col_Gfos_A, pch=16) points(Hs_Mod_B~meanHs_B_updist,col=col_Gfos_B, pch=16) for (j in 1:length(meanHs_A_red_updist)){ point <- cbind(meanHs_A_red_updist,Hs_Mod_A)[j,] seg <- unlist(perp.segment.coord(point[1],point[2])) segments(seg[1],seg[2],seg[3],seg[4], col=col_Gfos_A, lty=2, lwd=0.5) } for (k in 1:length(meanHs_B_updist)){ point <- cbind(meanHs_B_updist,Hs_Mod_B)[k,] seg <- unlist(perp.segment.coord(point[1],point[2])) segments(seg[1],seg[2],seg[3],seg[4], col=col_Gfos_B, lty=2, lwd=0.5) } textbox(max(Hs_modelled[,-1],meanHs_A_red_updist,meanHs_B_updist, na.rm=T),min(Hs_modelled[,-1],meanHs_A_red_updist,meanHs_B_updist, na.rm=T)+0.18,paste0(measure1_short," = ",formatC(round(sum_orthdist_Hs_A[[i-1]],sum_digits),digits=sum_digits, format="f")), txt.cex=sum_cex, txt.adj=1, txt.col=col_Gfos_A, frm.col=white_transparent, frm.brd = NA, frm.siz = 0.2) textbox(max(Hs_modelled[,-1],meanHs_A_red_updist,meanHs_B_updist, na.rm=T),min(Hs_modelled[,-1],meanHs_A_red_updist,meanHs_B_updist, na.rm=T)+0.05,paste0(measure1_short," = ",formatC(round(sum_orthdist_Hs_B[[i-1]],sum_digits),digits=sum_digits, format="f")), txt.cex=sum_cex, txt.adj=1, txt.col=col_Gfos_B, frm.col=white_transparent, frm.brd = NA, frm.siz = 0.2) } for (i in 1:length(lab)){ plot.new() if (i %in% c(1:2)){ text(0.5,1,lab[i],adj=c(0.5,1), cex=lab_cex[i]) } if (i %in% c(3:5)){ text(0,0.5,lab[i],adj=c(0,0.5), cex=lab_cex[i]) } } plot.new() legend("topleft",c(label_A,label_B),pch = 16, col = c(col_Gfos_A,col_Gfos_B), bty="n", cex=2) text(0,0,measure1, adj=0) dev.off() ##** TABLE S17 #### SPOrank_Hs_A <- labs_comb[order(apply(hist_Hs_A,2,sum))] SPOrank_Hs_B <- labs_comb[order(apply(hist_Hs_B,2,sum))] SPOrank_spo_Hs_A <- apply(hist_Hs_A,2,sum)[order(apply(hist_Hs_A,2,sum))] SPOrank_d_Hs_A <- labs_short[,3][order(apply(hist_Hs_A,2,sum))] SPOrank_W_Hs_A <- labs_short[,2][order(apply(hist_Hs_A,2,sum))] SPOrank_K_Hs_A <- labs_short[,1][order(apply(hist_Hs_A,2,sum))] SPOrank_spo_Hs_B <- apply(hist_Hs_B,2,sum)[order(apply(hist_Hs_B,2,sum))] SPOrank_d_Hs_B <- labs_short[,3][order(apply(hist_Hs_B,2,sum))] SPOrank_W_Hs_B <- labs_short[,2][order(apply(hist_Hs_B,2,sum))] SPOrank_K_Hs_B <- labs_short[,1][order(apply(hist_Hs_B,2,sum))] SPOrank_Hs <- data.frame("He_A_SPO"=SPOrank_spo_Hs_A,"He_A_d"=SPOrank_d_Hs_A,"He_A_W"=SPOrank_W_Hs_A,"He_A_K"=SPOrank_K_Hs_A, "He_B_SPO"=SPOrank_spo_Hs_B,"He_B_d"=SPOrank_d_Hs_B,"He_B_W"=SPOrank_W_Hs_B,"He_B_K"=SPOrank_K_Hs_B) write.csv2(SPOrank_Hs, paste0(WD,"/Analysis_",output,"/SuppFigs/Table_S17.csv")) ##** TABLE S18 #### MPOrank_Ar_A <- labs_comb[order(apply(hist_Ar_A,2,median))] MPOrank_Ar_B <- labs_comb[order(apply(hist_Ar_B,2,median))] MPOrank_mpo_Ar_A <- apply(hist_Ar_A,2,median)[order(apply(hist_Ar_A,2,median))] MPOrank_d_Ar_A <- labs_short[,3][order(apply(hist_Ar_A,2,median))] MPOrank_W_Ar_A <- labs_short[,2][order(apply(hist_Ar_A,2,median))] MPOrank_K_Ar_A <- labs_short[,1][order(apply(hist_Ar_A,2,median))] MPOrank_mpo_Ar_B <- apply(hist_Ar_B,2,median)[order(apply(hist_Ar_B,2,median))] MPOrank_d_Ar_B <- labs_short[,3][order(apply(hist_Ar_B,2,median))] MPOrank_W_Ar_B <- labs_short[,2][order(apply(hist_Ar_B,2,median))] MPOrank_K_Ar_B <- labs_short[,1][order(apply(hist_Ar_B,2,median))] MPOrank_Ar <- data.frame("Ar_A_MPO"=MPOrank_mpo_Ar_A,"Ar_A_d"=MPOrank_d_Ar_A,"Ar_A_W"=MPOrank_W_Ar_A,"Ar_A_K"=MPOrank_K_Ar_A, "Ar_B_MPO"=MPOrank_mpo_Ar_B,"Ar_B_d"=MPOrank_d_Ar_B,"Ar_B_W"=MPOrank_W_Ar_B,"Ar_B_K"=MPOrank_K_Ar_B) write.csv2(MPOrank_Ar, paste0(WD,"/Analysis_",output,"/SuppFigs/Table_S18.csv")) ##** TABLE S19 #### MPOrank_Ho_A <- labs_comb[order(apply(hist_Ho_A,2,median))] MPOrank_Ho_B <- labs_comb[order(apply(hist_Ho_B,2,median))] MPOrank_mpo_Ho_A <- apply(hist_Ho_A,2,median)[order(apply(hist_Ho_A,2,median))] MPOrank_d_Ho_A <- labs_short[,3][order(apply(hist_Ho_A,2,median))] MPOrank_W_Ho_A <- labs_short[,2][order(apply(hist_Ho_A,2,median))] MPOrank_K_Ho_A <- labs_short[,1][order(apply(hist_Ho_A,2,median))] MPOrank_mpo_Ho_B <- apply(hist_Ho_B,2,median)[order(apply(hist_Ho_B,2,median))] MPOrank_d_Ho_B <- labs_short[,3][order(apply(hist_Ho_B,2,median))] MPOrank_W_Ho_B <- labs_short[,2][order(apply(hist_Ho_B,2,median))] MPOrank_K_Ho_B <- labs_short[,1][order(apply(hist_Ho_B,2,median))] MPOrank_Ho <- data.frame("Ho_A_MPO"=MPOrank_mpo_Ho_A,"Ho_A_d"=MPOrank_d_Ho_A,"Ho_A_W"=MPOrank_W_Ho_A,"Ho_A_K"=MPOrank_K_Ho_A, "Ho_B_MPO"=MPOrank_mpo_Ho_B,"Ho_B_d"=MPOrank_d_Ho_B,"Ho_B_W"=MPOrank_W_Ho_B,"Ho_B_K"=MPOrank_K_Ho_B) write.csv2(MPOrank_Ho, paste0(WD,"/Analysis_",output,"/SuppFigs/Table_S19.csv")) ##** TABLE S20 #### MPOrank_Hs_A <- labs_comb[order(apply(hist_Hs_A,2,median))] MPOrank_Hs_B <- labs_comb[order(apply(hist_Hs_B,2,median))] MPOrank_mpo_Hs_A <- apply(hist_Hs_A,2,median)[order(apply(hist_Hs_A,2,median))] MPOrank_d_Hs_A <- labs_short[,3][order(apply(hist_Hs_A,2,median))] MPOrank_W_Hs_A <- labs_short[,2][order(apply(hist_Hs_A,2,median))] MPOrank_K_Hs_A <- labs_short[,1][order(apply(hist_Hs_A,2,median))] MPOrank_mpo_Hs_B <- apply(hist_Hs_B,2,median)[order(apply(hist_Hs_B,2,median))] MPOrank_d_Hs_B <- labs_short[,3][order(apply(hist_Hs_B,2,median))] MPOrank_W_Hs_B <- labs_short[,2][order(apply(hist_Hs_B,2,median))] MPOrank_K_Hs_B <- labs_short[,1][order(apply(hist_Hs_B,2,median))] MPOrank_Hs <- data.frame("He_A_MPO"=MPOrank_mpo_Hs_A,"He_A_d"=MPOrank_d_Hs_A,"He_A_W"=MPOrank_W_Hs_A,"He_A_K"=MPOrank_K_Hs_A, "He_B_MPO"=MPOrank_mpo_Hs_B,"He_B_d"=MPOrank_d_Hs_B,"He_B_W"=MPOrank_W_Hs_B,"He_B_K"=MPOrank_K_Hs_B) write.csv2(MPOrank_Hs, paste0(WD,"/Analysis_",output,"/SuppFigs/Table_S20.csv")) ##** FIG S21 #### ##** Histogram of directed perpendicular offset to 1:1 line main_s21=F mp_dim <- multipanel.dimensions(main.col=pla$a,main.row=pla$b,pla$x,pla$y,sub1=pla$sub1,sub2=pla$sub2,main=main_s21, h.main=0.5) if(pdf){ pdf(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS21.pdf"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1])) }else{ png(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS21.png"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1]), units="in", res=600) } nf <- multipanel.layout(main.col=pla$a,main.row=pla$b,pla$x,pla$y,sub1=pla$sub1,sub2=pla$sub2,main=main_s21, h.main=0.5) for (i in 1:ncol(hist_Ar_A)){ par(mar=c(0,0,0,0)) yclip <- 28 ylim <- 35 hist(hist_Ar_A_directed[,i], breaks=seq(floor(min(hist_Ar_A_directed,hist_Ar_B_directed, na.rm=T)),ceiling(max(hist_Ar_A_directed,hist_Ar_B_directed, na.rm=T))-0.5,0.5), xlim=c(floor(min(hist_Ar_A_directed,hist_Ar_B_directed, na.rm=T)),ceiling(max(hist_Ar_A_directed,hist_Ar_B_directed, na.rm=T))), ylim=c(0,ylim), xaxt="n", yaxt="n", col=col_Gfos_A, main="") clip(floor(min(hist_Ar_A_directed,hist_Ar_B_directed, na.rm=T)),ceiling(max(hist_Ar_A_directed,hist_Ar_B_directed, na.rm=T)), 0, yclip) abline(v=0, lwd=1.5, lty=2) abline(v=median(hist_Ar_A_directed[,i], na.rm=T),col=col_Gfos_A) abline(v=median(hist_Ar_B_directed[,i], na.rm=T),col=col_Gfos_B) clip(floor(min(hist_Ar_A_directed,hist_Ar_B_directed, na.rm=T)),ceiling(max(hist_Ar_A_directed,hist_Ar_B_directed, na.rm=T)), 0, ylim) textbox(ceiling(max(hist_Ar_A_directed,hist_Ar_B_directed, na.rm=T)),30,paste0(measure4," = ",formatC(round(median(hist_Ar_A_directed[,i], na.rm=T),median_digits),digits=median_digits, format="f")), txt.cex=median_cex, txt.col=col_Gfos_A, txt.adj=1, frm.col="white", frm.brd = "white", frm.siz = 0.2) textbox(ceiling(max(hist_Ar_A_directed,hist_Ar_B_directed, na.rm=T)),25,paste0(measure4," = ",formatC(round(median(hist_Ar_B_directed[,i], na.rm=T),median_digits),digits=median_digits, format="f")), txt.cex=median_cex, txt.col=col_Gfos_B, txt.adj=1, frm.col="white", frm.brd = "white", frm.siz = 0.2) hist(hist_Ar_A_directed[,i], breaks=seq(floor(min(hist_Ar_A_directed,hist_Ar_B_directed, na.rm=T)),ceiling(max(hist_Ar_A_directed,hist_Ar_B_directed, na.rm=T))-0.5,0.5), xlim=c(floor(min(hist_Ar_A_directed,hist_Ar_B_directed, na.rm=T)),ceiling(max(hist_Ar_A_directed,hist_Ar_B_directed, na.rm=T))), ylim=c(0,ylim), xaxt="n", yaxt="n", col=col_Gfos_A, main="", add=T) hist(hist_Ar_B_directed[,i], breaks=seq(floor(min(hist_Ar_A_directed,hist_Ar_B_directed, na.rm=T)),ceiling(max(hist_Ar_A_directed,hist_Ar_B_directed, na.rm=T))-0.5,0.5), xlim=c(floor(min(hist_Ar_A_directed,hist_Ar_B_directed, na.rm=T)),ceiling(max(hist_Ar_A_directed,hist_Ar_B_directed, na.rm=T))), ylim=c(0,ylim), xaxt="n", yaxt="n", col=col_Gfos_B, main="", add=T) if (i %in% c(1,2)){ mtext(lab[6], side=3, line=0.8, cex=1.3) } if (i %in% c(7,8)){ mtext(lab[7], side=3, line=0.8, cex=1.3) } if (i %in% c(13,14)){ mtext(lab[8], side=3, line=0.8, cex=1.3) } if (i %in% c(3)){ mtext("Frequencies [counts]", side=2, line=1, cex=1) } if (i %in% c(11,12)){ mtext(measure3, side=1, line=3, cex=1) } if (i %in% c(5:6,11:12,17:18)){ axis(1) } if (i %in% c(1,3,5)){ axis(2, at=seq(0,yclip,5), labels=F) } if (i %in% c(2,4,6)){ axis(2, at=seq(0,yclip,5)) } if (i %in% c(13:18)){ axis(4, at=seq(0,yclip,5), labels=F) } } if(main_s21){ plot.new() text(0.5,0.7,paste0(lab_Ar,": Distribution of directed ",measure3a),adj=c(0.5,0.5),cex=3) } for (i in 1:length(lab)){ plot.new() if (i %in% c(1:2)){ text(0.5,1,lab[i],adj=c(0.5,1), cex=lab_cex[i]) } if (i %in% c(3:5)){ text(0,0.5,lab[i],adj=c(0,0.5), cex=lab_cex[i]) } } par(mar=c(0,6,16,0.5), pty="s") i <- ncol(Ar_modelled) Ar_Mod_A <- Ar_modelled[,i][rownames(Ar_modelled)%in%modsite_GfosA] Ar_Mod_B <- Ar_modelled[,i][rownames(Ar_modelled)%in%modsite_GfosB] plot(Ar_Mod_A~meanAr_A_red_updist, type="n", xlim=c(min(Ar_modelled[,-1], na.rm=T),max(Ar_modelled[,-1], na.rm=T)), ylim=c(min(Ar_modelled[,-1], na.rm=T),max(Ar_modelled[,-1], na.rm=T)), xaxt="n", yaxt="n", xlab="", ylab="", bty="n", asp=1) par(xpd=T) legend(-5,yclip,c(label_A,label_B),pch = 16, col = c(col_Gfos_A,col_Gfos_B), bty="n", cex=2) par(xpd=F) dev.off() ##** FIG S22 #### ##** Histogram of directed perpendicular offset to 1:1 line main_s22=F mp_dim <- multipanel.dimensions(main.col=pla$a,main.row=pla$b,pla$x,pla$y,sub1=pla$sub1,sub2=pla$sub2,main=main_s22, h.main=0.5) if(pdf){ pdf(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS22.pdf"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1])) }else{ png(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS22.png"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1]), units="in", res=600) } nf <- multipanel.layout(main.col=pla$a,main.row=pla$b,pla$x,pla$y,sub1=pla$sub1,sub2=pla$sub2,main=main_s22, h.main=0.5) for (i in 1:ncol(hist_Ho_A_directed)){ par(mar=c(0,0,0,0)) yclip <- 30 ylim <- 35 # xlim <- round(max(abs(min(hist_Ho_A_directed,hist_Ho_B_directed)),max(hist_Ho_A_directed,hist_Ho_B_directed)),1) xlim <- ceiling(max(hist_Ho_A_directed,hist_Ho_B_directed, na.rm=T)/0.1)*0.1 hist(hist_Ho_A_directed[,i], breaks=seq(-xlim,xlim,0.05), xlim=c(-xlim,xlim), ylim=c(0,ylim), xaxt="n", yaxt="n", col=col_Gfos_A, main="") clip(-xlim,xlim, 0, yclip) abline(v=0, lwd=1.5, lty=2) abline(v=median(hist_Ho_A_directed[,i], na.rm=T),col=col_Gfos_A) abline(v=median(hist_Ho_B_directed[,i], na.rm=T),col=col_Gfos_B) clip(-xlim,xlim, 0, ylim) textbox(xlim,30,paste0(measure4," = ",formatC(round(median(hist_Ho_A_directed[,i], na.rm=T),median_digits),digits=median_digits, format="f")), txt.cex=median_cex, txt.col=col_Gfos_A, txt.adj=1, frm.col="white", frm.brd = "white", frm.siz = 0.2) textbox(xlim,25,paste0(measure4," = ",formatC(round(median(hist_Ho_B_directed[,i], na.rm=T),median_digits),digits=median_digits, format="f")), txt.cex=median_cex, txt.col=col_Gfos_B, txt.adj=1, frm.col="white", frm.brd = "white", frm.siz = 0.2) hist(hist_Ho_A_directed[,i], breaks=seq(-xlim,xlim,0.05), xlim=c(-xlim,xlim), ylim=c(0,ylim), xaxt="n", yaxt="n", col=col_Gfos_A, main="", add=T) hist(hist_Ho_B_directed[,i], breaks=seq(-xlim,xlim,0.05), xlim=c(-xlim,xlim), ylim=c(0,ylim), xaxt="n", yaxt="n", col=col_Gfos_B, main="", add=T) if (i %in% c(1,2)){ mtext(lab[6], side=3, line=0.8, cex=1.3) } if (i %in% c(7,8)){ mtext(lab[7], side=3, line=0.8, cex=1.3) } if (i %in% c(13,14)){ mtext(lab[8], side=3, line=0.8, cex=1.3) } if (i %in% c(3)){ mtext("Frequencies [counts]", side=2, line=1, cex=1) } if (i %in% c(11,12)){ mtext(measure3, side=1, line=3, cex=1) } if (i %in% c(5:6,11:12,17:18)){ axis(1, at=as.numeric(formatC(seq(-xlim+0.1,xlim-0.1,0.1),digits=1, format="f")), labels=T) } if (i %in% c(1,3,5)){ axis(2, at=seq(0,yclip,5), labels=F) } if (i %in% c(2,4,6)){ axis(2, at=seq(0,yclip,5)) } if (i %in% c(13:18)){ axis(4, at=seq(0,yclip,5), labels=F) } } if(main_s22){ plot.new() text(0.5,0.7,paste0(lab_Ho,": Distribution of directed ",measure3a),adj=c(0.5,0.5),cex=3) } for (i in 1:length(lab)){ plot.new() if (i %in% c(1:2)){ text(0.5,1,lab[i],adj=c(0.5,1), cex=lab_cex[i]) } if (i %in% c(3:5)){ text(0,0.5,lab[i],adj=c(0,0.5), cex=lab_cex[i]) } } par(mar=c(0,6,16,0.5), pty="s") i <- ncol(Ar_modelled) Ar_Mod_A <- Ar_modelled[,i][rownames(Ar_modelled)%in%modsite_GfosA] Ar_Mod_B <- Ar_modelled[,i][rownames(Ar_modelled)%in%modsite_GfosB] plot(Ar_Mod_A~meanAr_A_red_updist, type="n", xlim=c(min(Ar_modelled[,-1], na.rm=T),max(Ar_modelled[,-1], na.rm=T)), ylim=c(min(Ar_modelled[,-1], na.rm=T),max(Ar_modelled[,-1], na.rm=T)), xaxt="n", yaxt="n", xlab="", ylab="", bty="n", asp=1) par(xpd=T) legend(-5,yclip,c(label_A,label_B),pch = 16, col = c(col_Gfos_A,col_Gfos_B), bty="n", cex=2) par(xpd=F) dev.off() ##** FIG S23 #### ##** Histogram of directed perpendicular offset to 1:1 line main_s23=F mp_dim <- multipanel.dimensions(main.col=pla$a,main.row=pla$b,pla$x,pla$y,sub1=pla$sub1,sub2=pla$sub2,main=main_s23, h.main=0.5) if(pdf){ pdf(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS23.pdf"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1])) }else{ png(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS23.png"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1]), units="in", res=600) } nf <- multipanel.layout(main.col=pla$a,main.row=pla$b,pla$x,pla$y,sub1=pla$sub1,sub2=pla$sub2,main=main_s23, h.main=0.5) for (i in 1:ncol(hist_Hs_A_directed)){ par(mar=c(0,0,0,0)) yclip <- 30 ylim <- 35 # xlim <- round(max(abs(min(hist_Hs_A_directed,hist_Hs_B_directed)),max(hist_Hs_A_directed,hist_Hs_B_directed)),1) xlim <- ceiling(max(hist_Hs_A_directed,hist_Hs_B_directed, na.rm=T)/0.1)*0.1 hist(hist_Hs_A_directed[,i], breaks=seq(-xlim,xlim,0.05), xlim=c(-xlim,xlim), ylim=c(0,ylim), xaxt="n", yaxt="n", col=col_Gfos_A, main="") clip(-xlim,xlim, 0, yclip) abline(v=0, lwd=1.5, lty=2) abline(v=median(hist_Hs_A_directed[,i], na.rm=T),col=col_Gfos_A) abline(v=median(hist_Hs_B_directed[,i], na.rm=T),col=col_Gfos_B) clip(-xlim,xlim, 0, ylim) textbox(xlim,30,paste0(measure4," = ",formatC(round(median(hist_Hs_A_directed[,i], na.rm=T),median_digits),digits=median_digits, format="f")), txt.cex=median_cex, txt.col=col_Gfos_A, txt.adj=1, frm.col="white", frm.brd = "white", frm.siz = 0.2) textbox(xlim,25,paste0(measure4," = ",formatC(round(median(hist_Hs_B_directed[,i], na.rm=T),median_digits),digits=median_digits, format="f")), txt.cex=median_cex, txt.col=col_Gfos_B, txt.adj=1, frm.col="white", frm.brd = "white", frm.siz = 0.2) hist(hist_Hs_A_directed[,i], breaks=seq(-xlim,xlim,0.05), xlim=c(-xlim,xlim), ylim=c(0,ylim), xaxt="n", yaxt="n", col=col_Gfos_A, main="", add=T) hist(hist_Hs_B_directed[,i], breaks=seq(-xlim,xlim,0.05), xlim=c(-xlim,xlim), ylim=c(0,ylim), xaxt="n", yaxt="n", col=col_Gfos_B, main="", add=T) if (i %in% c(1,2)){ mtext(lab[6], side=3, line=0.8, cex=1.3) } if (i %in% c(7,8)){ mtext(lab[7], side=3, line=0.8, cex=1.3) } if (i %in% c(13,14)){ mtext(lab[8], side=3, line=0.8, cex=1.3) } if (i %in% c(3)){ mtext("Frequencies [counts]", side=2, line=1, cex=1) } if (i %in% c(11,12)){ mtext(measure3, side=1, line=3, cex=1) } if (i %in% c(5:6,11:12,17:18)){ axis(1, at=as.numeric(formatC(seq(-xlim+0.1,xlim-0.1,0.1),digits=1, format="f")), labels=T) } if (i %in% c(1,3,5)){ axis(2, at=seq(0,yclip,5), labels=F) } if (i %in% c(2,4,6)){ axis(2, at=seq(0,yclip,5)) } if (i %in% c(13:18)){ axis(4, at=seq(0,yclip,5), labels=F) } } if(main_s23){ plot.new() text(0.5,0.7,paste0(lab_Hs,": Distribution of directed ",measure3a),adj=c(0.5,0.5),cex=3) } for (i in 1:length(lab)){ plot.new() if (i %in% c(1:2)){ text(0.5,1,lab[i],adj=c(0.5,1), cex=lab_cex[i]) } if (i %in% c(3:5)){ text(0,0.5,lab[i],adj=c(0,0.5), cex=lab_cex[i]) } } par(mar=c(0,6,16,0.5), pty="s") i <- ncol(Ar_modelled) Ar_Mod_A <- Ar_modelled[,i][rownames(Ar_modelled)%in%modsite_GfosA] Ar_Mod_B <- Ar_modelled[,i][rownames(Ar_modelled)%in%modsite_GfosB] plot(Ar_Mod_A~meanAr_A_red_updist, type="n", xlim=c(min(Ar_modelled[,-1], na.rm=T),max(Ar_modelled[,-1], na.rm=T)), ylim=c(min(Ar_modelled[,-1], na.rm=T),max(Ar_modelled[,-1], na.rm=T)), xaxt="n", yaxt="n", xlab="", ylab="", bty="n", asp=1) par(xpd=T) legend(-5,yclip,c(label_A,label_B),pch = 16, col = c(col_Gfos_A,col_Gfos_B), bty="n", cex=2) par(xpd=F) dev.off() if(fst){ ##** FIG S24 #### # Fst perpendicular offset ##*** Preparing matrices#### orthdist_Fst_A <- vector() orthdist_Fst_A_directed <- vector() sum_orthdist_Fst_A <- vector() hist_Fst_A <- matrix(nrow=length(fst_A_red), ncol=ncol(Ar_modelled)-1) hist_Fst_A_directed <- matrix(nrow=length(fst_A_red), ncol=ncol(Ar_modelled)-1) orthdist_Fst_B <- vector() orthdist_Fst_B_directed <- vector() sum_orthdist_Fst_B <- vector() hist_Fst_B <- matrix(nrow=length(fst_B), ncol=ncol(Ar_modelled)-1) hist_Fst_B_directed <- matrix(nrow=length(fst_B), ncol=ncol(Ar_modelled)-1) # dir.create(paste0(WD,"/Analysis_",output,"/SuppFigs/Fst"), showWarnings=F) r <- 0 for (d in 1:length(D)){ # looping over dispersal rates for (w in 1:length(W)){ # looping over dispersal directionalities for (k in 1:length(K)){ # looping over carrying capacities r <- r+1 if(existing_data){ load(paste0(DF,"/02_Data_prep/Fst_data/FstData_",D[d],"_",W[w],"_",K[k],".Rdata")) }else{ load(paste0(DF,"/02_Data_prep/",prep_folder,"/IndPopGenData_",D[d],"_",W[w],"_",K[k],".Rdata")) } # Distance matrix to vector MEANFST_Mod <- meanFst_Mod rownames(MEANFST_Mod) <- modsite colnames(MEANFST_Mod) <- modsite fst_match_B <- match(modsite[modsite%in%microsite_B],microsite_B) MEANFST_Mod_B <- MEANFST_Mod[fst_match_B,fst_match_B] meanFst_Mod_B <- MEANFST_Mod_B[upper.tri(MEANFST_Mod_B)] fst_match_A_red <- match(modsite[modsite%in%microsite_A],microsite_A_red) MEANFST_Mod_A_red <- MEANFST_Mod[fst_match_A_red,fst_match_A_red] meanFst_Mod_A_red <- MEANFST_Mod_A_red[upper.tri(MEANFST_Mod_A_red)] seg <- matrix(nrow=length(fst_A_red),ncol=4) point <- cbind(fst_A_red,meanFst_Mod_A_red) for (j in 1:nrow(point)){ seg[j,] <- unlist(perp.segment.coord(point[j,1],point[j,2])) orthdist_Fst_A[j] <- euc.dist(c(seg[j,1],seg[j,2]),c(seg[j,3],seg[j,4])) orthdist_Fst_A_directed[j] <- sign(seg[j,2]-seg[j,4])*orthdist_Fst_A[j] } seg <- matrix(nrow=length(fst_B),ncol=4) point <- cbind(fst_B,meanFst_Mod_B) for (j in 1:nrow(point)){ seg[j,] <- unlist(perp.segment.coord(point[j,1],point[j,2])) orthdist_Fst_B[j] <- euc.dist(c(seg[j,1],seg[j,2]),c(seg[j,3],seg[j,4])) orthdist_Fst_B_directed[j] <- sign(seg[j,2]-seg[j,4])*orthdist_Fst_B[j] } sum_orthdist_Fst_A[[r]] <- sum(orthdist_Fst_A) sum_orthdist_Fst_B[[r]] <- sum(orthdist_Fst_B) hist_Fst_A[,r] <- orthdist_Fst_A hist_Fst_B[,r] <- orthdist_Fst_B hist_Fst_A_directed[,r] <- orthdist_Fst_A_directed hist_Fst_B_directed[,r] <- orthdist_Fst_B_directed label_Mod_short <- paste0("D",D_label[d],"_W",W_label[w],"_K",K_label[k]) # if(pdf){ # pdf(paste0(WD,"/Analysis_",output,"/SuppFigs/Fst/Fig_FstMod_",label_Mod_short,".pdf"), width=6, height=6) # }else{ # png(paste0(WD,"/Analysis_",output,"/SuppFigs/Fst/Fig_FstMod_",label_Mod_short,".png"), width=6, height=6, units="in", res=300) # } plot(fst_A_red,meanFst_Mod_A_red, xlim=c(min(fst_A_red,meanFst_Mod_A_red,fst_B,meanFst_Mod_B, na.rm=T), max(fst_A_red,meanFst_Mod_A_red,fst_B,meanFst_Mod_B, na.rm=T)), ylim=c(min(fst_A_red,meanFst_Mod_A_red,fst_B,meanFst_Mod_B, na.rm=T), max(fst_A_red,meanFst_Mod_A_red,fst_B,meanFst_Mod_B, na.rm=T)), col=col_Gfos_A, asp=1) points(fst_B,meanFst_Mod_B, col=col_Gfos_B) abline(0,1,col="red") mtext(label_Mod_short) # dev.off() } # end looping over carrying capacities } # end looping over dispersal directionalities } # end looping over dispersal rates names(sum_orthdist_Fst_A) <- colnames(Ar_modelled)[-1] names(sum_orthdist_Fst_B) <- colnames(Ar_modelled)[-1] #### Spatial distance between populations in simulations mod_vertices <- match(modsite,V(net)$name) DIST_Mod <- distances(net, v=V(net)[mod_vertices], to=V(net)[mod_vertices], weights=E(net)) #### Distance matrix to vector dist_Mod <- DIST_Mod[upper.tri(DIST_Mod)] mp_dim <- multipanel.dimensions(main.col=pla$a,main.row=pla$b,pla$x,pla$y,sub1=pla$sub1,sub2=pla$sub2,main=pla$main) if(pdf){ pdf(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS24.pdf"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1])) }else{ png(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS24.png"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1]), units="in", res=600) } nf <- multipanel.layout(main.col=pla$a,main.row=pla$b,pla$x,pla$y,sub1=pla$sub1,sub2=pla$sub2,main=pla$main) r <- 0 for (d in 1:length(D)){ # looping over dispersal rates for (w in 1:length(W)){ # looping over dispersal directionalities for (k in 1:length(K)){ # looping over carrying capacities r <- r+1 if(existing_data){ load(paste0(DF,"/02_Data_prep/Fst_data/FstData_",D[d],"_",W[w],"_",K[k],".Rdata")) }else{ load(paste0(DF,"/02_Data_prep/",prep_folder,"/IndPopGenData_",D[d],"_",W[w],"_",K[k],".Rdata")) } # Distance matrix to vector MEANFST_Mod <- meanFst_Mod meanFst_Mod <- meanFst_Mod[upper.tri(meanFst_Mod)] # DISTDATA table construction DISTDATA_Mod <- cbind(meanFst_Mod,dist_Mod) colnames(DISTDATA_Mod) <- c("fst","dist") DISTDATA_Mod <- data.frame(DISTDATA_Mod) # Order DISTDATA according to dist DISTDATA_Mod <- DISTDATA_Mod[ order(DISTDATA_Mod$dist), ] # Remove NAs DISTDATA_Mod <- DISTDATA_Mod[which(!is.na(DISTDATA_Mod$fst)),] # Prepare DISTDATA with non-zero Fst values DISTDATA_Mod$nonneg_fst <- DISTDATA_Mod$fst DISTDATA_Mod$nonneg_fst[which(DISTDATA_Mod$fst<0)] <- 0 # Prepare DISTDATA with log-transformed dist values DISTDATA_Mod$log_dist <- log(DISTDATA_Mod$dist) #### Combined GLM of genetic differentiation by instream distance * species (power term) power <- seq(0,1,0.01) AICpower <- c() for (i in 1:length(power)){ pow.mod.Fst <- lm(nonneg_fst ~ I(dist^power[i]), DISTDATA_Mod, na.action = "na.fail") AICpower[i] <- AIC(pow.mod.Fst) } model <- lm.bind(nonneg_fst ~ I(dist^power[which.min(AICpower)]), DISTDATA_Mod, "fst_power", critval, step=T) DISTDATA_Mod <- model[[1]] lm_fst_power <- model[[2]] slm_fst_power <- model[[3]] DISTDATA_Mod$spec <- "Mod" DISTDATA_Mod <- DISTDATA_Mod[,-which(colnames(DISTDATA_Mod)=="log_dist")] DISTDATA_temp <- DISTDATA[,match(colnames(DISTDATA_Mod),colnames(DISTDATA))] DISTDATA_temp <- rbind(DISTDATA_Mod,DISTDATA_temp) label_Mod <- paste0("D=",D_label[d],", W_up=",W_label[w],", K=",K_label[k]) label_Mod_short <- paste0("D",D_label[d],"_W",W_label[w],"_K",K_label[k]) par(mar=c(0,0,0,0)) x <- "fst" y <- "dist" dat=DISTDATA_temp model="slm_fst_power" CI_border = F xlabel="Instream distance [km]" ylabel=expression('Genetic diff. [Pairwise Nei F'[ST]*']') xax="n" yax="n" bty="o" yrange=c(0,1) pointtrans = T trans=0.2 trans_mod=0.1 xrev=F axislog="" pt.cex=0.5 lwd=1 cex.lab=2 cex.axis=1.5 legend=F cex.legend=1.5 main=F col1=col_Gfos_A col2=col_Gfos_B col3=colMod col1trans <- rgb(col2rgb(col1)[1,]/255,col2rgb(col1)[2,]/255,col2rgb(col1)[3,]/255,trans) col2trans <- rgb(col2rgb(col2)[1,]/255,col2rgb(col2)[2,]/255,col2rgb(col2)[3,]/255,trans) col3trans <- rgb(col2rgb(col3)[1,]/255,col2rgb(col3)[2,]/255,col2rgb(col3)[3,]/255,trans_mod) form <- reformulate(y, response = x) formfit <- reformulate(y, response = paste0(model,"_fit")) formupr <- reformulate(y, response = paste0(model,"_upr")) formlwr <- reformulate(y, response = paste0(model,"_lwr")) xcol <- which(colnames(dat)==x) ycol <- which(colnames(dat)==y) lwrcol <- which(colnames(dat)==paste0(model,"_lwr")) uprcol <- which(colnames(dat)==paste0(model,"_upr")) DATAordered <- dat[order(dat[,ycol]),] left <- min(DATAordered[,ycol], na.rm=T) right <- max(DATAordered[,ycol], na.rm=T) if (xrev==T){ xrange <- c(right,left) }else{ xrange <- c(left,right) } if (pointtrans==T){ col1point <- col1trans col2point <- col2trans col3point <- col3trans }else{ col1point <- col1 col2point <- col2 col3point <- col3 } plot(form, dat, type = "n", las = 1, bty = bty, xlab=xlabel, ylab=ylabel, xlim=xrange, ylim=yrange, log=axislog, xaxt=xax, yaxt=yax, cex.lab=cex.lab, cex.axis=cex.axis) polygon(c(rev(DATAordered[,ycol][DATAordered$spec=="A"]), DATAordered[,ycol][DATAordered$spec=="A"]), c(rev(DATAordered[,lwrcol][DATAordered$spec=="A"]), DATAordered[,uprcol][DATAordered$spec=="A"]), col = col1trans, border = NA) polygon(c(rev(DATAordered[,ycol][DATAordered$spec=="B"]), DATAordered[,ycol][DATAordered$spec=="B"]), c(rev(DATAordered[,lwrcol][DATAordered$spec=="B"]), DATAordered[,uprcol][DATAordered$spec=="B"]), col = col2trans, border = NA) polygon(c(rev(DATAordered[,ycol][DATAordered$spec=="Mod"]), DATAordered[,ycol][DATAordered$spec=="Mod"]), c(rev(DATAordered[,lwrcol][DATAordered$spec=="Mod"]), DATAordered[,uprcol][DATAordered$spec=="Mod"]), col = col3trans, border = NA) points(form, data = subset(DATAordered, spec == "A"), pch = 16, col = col1point, cex=pt.cex) points(form, data = subset(DATAordered, spec == "B"), pch = 16, col = col2point, cex=pt.cex) points(form, data = subset(DATAordered, spec == "Mod"), pch = 16, col = col3point, cex=pt.cex) lines(formfit, data = subset(DATAordered, spec == "A"), lwd = lwd, col=col1) if(CI_border){lines(formupr, data = subset(DATAordered, spec == "A"), lwd = 2, lty=2, col=col1)} if(CI_border){lines(formlwr, data = subset(DATAordered, spec == "A"), lwd = 2, lty=2, col=col1)} lines(formfit, data = subset(DATAordered, spec == "B"), lwd = lwd, col=col2) if(CI_border){lines(formupr, data = subset(DATAordered, spec == "B"), lwd = 2, lty=2, col=col2)} if(CI_border){lines(formlwr, data = subset(DATAordered, spec == "B"), lwd = 2, lty=2, col=col2)} lines(formfit, data = subset(DATAordered, spec == "Mod"), lwd = lwd, col=col3) if(CI_border){lines(formupr, data = subset(DATAordered, spec == "Mod"), lwd = 2, lty=2, col=col3)} if(CI_border){lines(formlwr, data = subset(DATAordered, spec == "Mod"), lwd = 2, lty=2, col=col3)} if (r%in%c(2,4,6)){ axis(2, cex.axis=1.5) } if (r%in%c(5,6,11,12,17,18)){ axis(1, c(0,50,100,150,200,250), at=c(0,50000,100000,150000,200000,250000), cex.axis=1.5) } if (r %in% c(13,15,17)){ axis(4, labels=F) } if (r %in% c(1,3,5)){ axis(2, labels=F) } if (r %in% c(1,2)){ mtext(lab[6], side=3, line=0.8, cex=1.4) } if (r %in% c(7,8)){ mtext(lab[7], side=3, line=0.8, cex=1.4) } if (r %in% c(13,14)){ mtext(lab[8], side=3, line=0.8, cex=1.4) } if (r %in% c(3)){ mtext(ylabel, side=2, line=1, cex=1) } if (r %in% c(11,12)){ mtext(xlabel, side=1, line=3, cex=1) } } # end looping over carrying capacities } # end looping over dispersal directionalities } # end looping over dispersal rates for (i in 1:length(lab)){ plot.new() if (i %in% c(1:2)){ text(0.5,1,lab[i],adj=c(0.5,1), cex=lab_cex[i]) } if (i %in% c(3:5)){ text(0,0.5,lab[i],adj=c(0,0.5), cex=lab_cex[i]) } } plot.new() legend("topleft",c(label_A,label_B, label_mod),pch = 16, col = c(col_Gfos_A,col_Gfos_B, colMod), bty="n", cex=2) dev.off() ##** FIG S25 #### ##** Histogram of directed perpendicular offset to 1:1 line main_s25=F median_cex <- 0.7 mp_dim <- multipanel.dimensions(main.col=pla$a,main.row=pla$b,pla$x,pla$y,sub1=pla$sub1,sub2=pla$sub2,main=main_s25, h.main=0.5) if(pdf){ pdf(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS25.pdf"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1])) }else{ png(paste0(WD,"/Analysis_",output,"/SuppFigs/FigS25.png"), width=fig.width, height=fig.width*(mp_dim[2]/mp_dim[1]), units="in", res=600) } nf <- multipanel.layout(main.col=pla$a,main.row=pla$b,pla$x,pla$y,sub1=pla$sub1,sub2=pla$sub2,main=main_s25, h.main=0.5) # min(abs(apply(hist_Fst_A_directed, 2, median))) for (i in 1:ncol(hist_Fst_A)){ par(mar=c(0,0,0,0)) yclip <- 450 ylim <- 500 hist(hist_Fst_A_directed[,i], breaks=seq(floor(min(hist_Fst_A_directed,hist_Fst_B_directed, na.rm=T)),ceiling(max(hist_Fst_A_directed,hist_Fst_B_directed, na.rm=T)),0.05), xlim=c(floor(min(hist_Fst_A_directed,hist_Fst_B_directed, na.rm=T)),ceiling(max(hist_Fst_A_directed,hist_Fst_B_directed, na.rm=T))), ylim=c(0,ylim), xaxt="n", yaxt="n", col=col_Gfos_A, main="") clip(floor(min(hist_Fst_A_directed,hist_Fst_B_directed, na.rm=T)),ceiling(max(hist_Fst_A_directed,hist_Fst_B_directed, na.rm=T)), 0, yclip) abline(v=0, lwd=1.5, lty=2) abline(v=median(hist_Fst_A_directed[,i]),col=col_Gfos_A) abline(v=median(hist_Fst_B_directed[,i]),col=col_Gfos_B) clip(floor(min(hist_Fst_A_directed,hist_Fst_B_directed, na.rm=T)),ceiling(max(hist_Fst_A_directed,hist_Fst_B_directed, na.rm=T)), 0, ylim) # textbox(ceiling(max(hist_Fst_A_directed,hist_Fst_B_directed, na.rm=T)),150,paste0(measure4," = ",formatC(round(median(hist_Fst_A_directed[,i]),median_digits),digits=median_digits, format="f")), txt.cex=median_cex, txt.col=col_Gfos_A, txt.adj=1, frm.col="white", frm.brd = "white", frm.siz = 0.2) # textbox(ceiling(max(hist_Fst_A_directed,hist_Fst_B_directed, na.rm=T)),50,paste0(measure4," = ",formatC(round(median(hist_Fst_B_directed[,i]),median_digits),digits=median_digits, format="f")), txt.cex=median_cex, txt.col=col_Gfos_B, txt.adj=1, frm.col="white", frm.brd = "white", frm.siz = 0.2) hist(hist_Fst_A_directed[,i], breaks=seq(floor(min(hist_Fst_A_directed,hist_Fst_B_directed, na.rm=T)),ceiling(max(hist_Fst_A_directed,hist_Fst_B_directed, na.rm=T)),0.05), xlim=c(floor(min(hist_Fst_A_directed,hist_Fst_B_directed, na.rm=T)),ceiling(max(hist_Fst_A_directed,hist_Fst_B_directed, na.rm=T))), ylim=c(0,ylim), xaxt="n", yaxt="n", col=col_Gfos_A, main="", add=T) hist(hist_Fst_B_directed[,i], breaks=seq(floor(min(hist_Fst_A_directed,hist_Fst_B_directed, na.rm=T)),ceiling(max(hist_Fst_A_directed,hist_Fst_B_directed, na.rm=T)),0.05), xlim=c(floor(min(hist_Fst_A_directed,hist_Fst_B_directed, na.rm=T)),ceiling(max(hist_Fst_A_directed,hist_Fst_B_directed, na.rm=T))), ylim=c(0,ylim), xaxt="n", yaxt="n", col=col_Gfos_B, main="", add=T) textbox(ceiling(max(hist_Fst_A_directed,hist_Fst_B_directed, na.rm=T)),yclip,paste0(measure4," = ",formatC(round(median(hist_Fst_A_directed[,i]),median_digits),digits=median_digits, format="f")), txt.cex=median_cex, txt.col=col_Gfos_A, txt.adj=1, frm.col="white", frm.brd = "white", frm.siz = 0.2) textbox(ceiling(max(hist_Fst_A_directed,hist_Fst_B_directed, na.rm=T)),yclip-50,paste0(measure4," = ",formatC(round(median(hist_Fst_B_directed[,i]),median_digits),digits=median_digits, format="f")), txt.cex=median_cex, txt.col=col_Gfos_B, txt.adj=1, frm.col="white", frm.brd = "white", frm.siz = 0.2) if (i %in% c(1,2)){ mtext(lab[6], side=3, line=0.8, cex=1.3) } if (i %in% c(7,8)){ mtext(lab[7], side=3, line=0.8, cex=1.3) } if (i %in% c(13,14)){ mtext(lab[8], side=3, line=0.8, cex=1.3) } if (i %in% c(3)){ mtext("Frequencies [counts]", side=2, line=1, cex=1) } if (i %in% c(11,12)){ mtext(measure3, side=1, line=3, cex=1) } if (i %in% c(5:6,11:12,17:18)){ axis(1) } '%notin%' <- Negate('%in%') if (i %notin% c(5:6,11:12,17:18)){ axis(1, labels = F) } if (i %in% c(1,3,5)){ axis(2, at=seq(0,yclip,200), labels=F) } if (i %in% c(2,4,6)){ axis(2, at=seq(0,yclip,200)) } if (i %in% c(13:18)){ axis(4, at=seq(0,yclip,200), labels=F) } } if(main_s25){ plot.new() text(0.5,0.7,paste0(lab_Ar,": Distribution of directed ",measure3a),adj=c(0.5,0.5),cex=3) } for (i in 1:length(lab)){ plot.new() if (i %in% c(1:2)){ text(0.5,1,lab[i],adj=c(0.5,1), cex=lab_cex[i]) } if (i %in% c(3:5)){ text(0,0.5,lab[i],adj=c(0,0.5), cex=lab_cex[i]) } } plot.new() legend("topleft",c(label_A,label_B),pch = 16, col = c(col_Gfos_A,col_Gfos_B), bty="n", cex=2) dev.off() ##** TABLE S26 #### SPOrank_Fst_A <- labs_comb[order(apply(hist_Fst_A,2,sum))] SPOrank_Fst_B <- labs_comb[order(apply(hist_Fst_B,2,sum))] SPOrank_spo_Fst_A <- apply(hist_Fst_A,2,sum)[order(apply(hist_Fst_A,2,sum))] SPOrank_d_Fst_A <- labs_short[,3][order(apply(hist_Fst_A,2,sum))] SPOrank_W_Fst_A <- labs_short[,2][order(apply(hist_Fst_A,2,sum))] SPOrank_K_Fst_A <- labs_short[,1][order(apply(hist_Fst_A,2,sum))] SPOrank_spo_Fst_B <- apply(hist_Fst_B,2,sum)[order(apply(hist_Fst_B,2,sum))] SPOrank_d_Fst_B <- labs_short[,3][order(apply(hist_Fst_B,2,sum))] SPOrank_W_Fst_B <- labs_short[,2][order(apply(hist_Fst_B,2,sum))] SPOrank_K_Fst_B <- labs_short[,1][order(apply(hist_Fst_B,2,sum))] SPOrank_Fst <- data.frame("Fst_A_SPO"=SPOrank_spo_Fst_A,"Fst_A_d"=SPOrank_d_Fst_A,"Fst_A_W"=SPOrank_W_Fst_A,"Fst_A_K"=SPOrank_K_Fst_A, "Fst_B_SPO"=SPOrank_spo_Fst_B,"Fst_B_d"=SPOrank_d_Fst_B,"Fst_B_W"=SPOrank_W_Fst_B,"Fst_B_K"=SPOrank_K_Fst_B) write.csv2(SPOrank_Fst, paste0(WD,"/Analysis_",output,"/SuppFigs/Table_S26.csv")) ##** TABLE S27 #### MPOrank_Fst_A <- labs_comb[order(apply(hist_Fst_A,2,median))] MPOrank_Fst_B <- labs_comb[order(apply(hist_Fst_B,2,median))] MPOrank_mpo_Fst_A <- apply(hist_Fst_A,2,median)[order(apply(hist_Fst_A,2,median))] MPOrank_d_Fst_A <- labs_short[,3][order(apply(hist_Fst_A,2,median))] MPOrank_W_Fst_A <- labs_short[,2][order(apply(hist_Fst_A,2,median))] MPOrank_K_Fst_A <- labs_short[,1][order(apply(hist_Fst_A,2,median))] MPOrank_mpo_Fst_B <- apply(hist_Fst_B,2,median)[order(apply(hist_Fst_B,2,median))] MPOrank_d_Fst_B <- labs_short[,3][order(apply(hist_Fst_B,2,median))] MPOrank_W_Fst_B <- labs_short[,2][order(apply(hist_Fst_B,2,median))] MPOrank_K_Fst_B <- labs_short[,1][order(apply(hist_Fst_B,2,median))] MPOrank_Fst <- data.frame("Fst_A_MPO"=MPOrank_mpo_Fst_A,"Fst_A_d"=MPOrank_d_Fst_A,"Fst_A_W"=MPOrank_W_Fst_A,"Fst_A_K"=MPOrank_K_Fst_A, "Fst_B_MPO"=MPOrank_mpo_Fst_B,"Fst_B_d"=MPOrank_d_Fst_B,"Fst_B_W"=MPOrank_W_Fst_B,"Fst_B_K"=MPOrank_K_Fst_B) write.csv2(MPOrank_Fst, paste0(WD,"/Analysis_",output,"/SuppFigs/Table_S27.csv")) ##** d: Model performance #### d0001_Fst_A <- hist_Fst_A[,(1:6)] d001_Fst_A <- hist_Fst_A[,c(7:12)] d01_Fst_A <- hist_Fst_A[,c(13:18)] d0001_Fst_B <- hist_Fst_B[,(1:6)] d001_Fst_B <- hist_Fst_B[,c(7:12)] d01_Fst_B <- hist_Fst_B[,c(13:18)] # Here we calculate the difference to d=0.01 (instead of d=0.001) diffSPO_d001both_Fst_A <- c(apply(d001_Fst_A,2,sum)-apply(d0001_Fst_A,2,sum),apply(d001_Fst_A,2,sum)-apply(d01_Fst_A,2,sum)) diffSPO_d001both_Fst_B <- c(apply(d001_Fst_B,2,sum)-apply(d0001_Fst_B,2,sum),apply(d001_Fst_B,2,sum)-apply(d01_Fst_B,2,sum)) diffMPO_d001both_Fst_A <- c(apply(d001_Fst_A,2,median)-apply(d0001_Fst_A,2,median),apply(d001_Fst_A,2,median)-apply(d01_Fst_A,2,median)) diffMPO_d001both_Fst_B <- c(apply(d001_Fst_B,2,median)-apply(d0001_Fst_B,2,median),apply(d001_Fst_B,2,median)-apply(d01_Fst_B,2,median)) list_d001both <- c(diffMPO_d001both_Fst_A,diffMPO_d001both_Fst_B, diffSPO_d001both_Fst_A,diffSPO_d001both_Fst_B) total_comparison_d001both <- length(na.omit(list_d001both)) improved_fit_d001both <- sum(na.omit(list_d001both<0)) # Model performance d improved_fit_d001both/total_comparison_d001both ##** W: Model performance #### w00_Fst_A <- hist_Fst_A[,c(1,2,7,8,13,14)] w05_Fst_A <- hist_Fst_A[,c(3,4,9,10,15,16)] w10_Fst_A <- hist_Fst_A[,c(5,6,11,12,17,18)] w00_Fst_B <- hist_Fst_B[,c(1,2,7,8,13,14)] w05_Fst_B <- hist_Fst_B[,c(3,4,9,10,15,16)] w10_Fst_B <- hist_Fst_B[,c(5,6,11,12,17,18)] diffSPO_w00both_Fst_A <- c(apply(w00_Fst_A,2,sum)-apply(w05_Fst_A,2,sum),apply(w00_Fst_A,2,sum)-apply(w10_Fst_A,2,sum)) diffSPO_w00both_Fst_B <- c(apply(w00_Fst_B,2,sum)-apply(w05_Fst_B,2,sum),apply(w00_Fst_B,2,sum)-apply(w10_Fst_B,2,sum)) diffMPO_w00both_Fst_A <- c(apply(w00_Fst_A,2,median)-apply(w05_Fst_A,2,median),apply(w00_Fst_A,2,median)-apply(w10_Fst_A,2,median)) diffMPO_w00both_Fst_B <- c(apply(w00_Fst_B,2,median)-apply(w05_Fst_B,2,median),apply(w00_Fst_B,2,median)-apply(w10_Fst_B,2,median)) list_w00both <- c(diffMPO_w00both_Fst_A,diffMPO_w00both_Fst_B, diffSPO_w00both_Fst_A,diffSPO_w00both_Fst_B) total_comparison_w00both <- length(na.omit(list_w00both)) improved_fit_w00both <- sum(na.omit(list_w00both<0)) # Model performance W improved_fit_w00both/total_comparison_w00both ##** K: Model performance #### k0_Fst_A <- hist_Fst_A[,c(1,3,5,7,9,11,13,15,17)] k1_Fst_A <- hist_Fst_A[,c(2,4,6,8,10,12,14,16,18)] k0_Fst_B <- hist_Fst_B[,c(1,3,5,7,9,11,13,15,17)] k1_Fst_B <- hist_Fst_B[,c(2,4,6,8,10,12,14,16,18)] diffSPO_k0k1_Fst_A <- apply(k0_Fst_A,2,sum)-apply(k1_Fst_A,2,sum) diffSPO_k0k1_Fst_B <- apply(k0_Fst_B,2,sum)-apply(k1_Fst_B,2,sum) diffMPO_k0k1_Fst_A <- apply(k0_Fst_A,2,median)-apply(k1_Fst_A,2,median) diffMPO_k0k1_Fst_B <- apply(k0_Fst_B,2,median)-apply(k1_Fst_B,2,median) list_k0k1 <- c(diffMPO_k0k1_Fst_A,diffMPO_k0k1_Fst_B, diffSPO_k0k1_Fst_A,diffSPO_k0k1_Fst_B) total_comparison_k0k1 <- length(na.omit(list_k0k1)) improved_fit_k0k1 <- sum(na.omit(list_k0k1)<0) # Model performance K improved_fit_k0k1/total_comparison_k0k1 }
6a3b49759e6a4a346a1a7e72529059671c2b40c3
c504cf54212655142112685e917904d41dfe0bde
/do_not_include/new.R
9f463f66ed45d4d4afc82214a07edbb429f27579
[]
no_license
BorisHouenou/WeightIt
3471ac96b4f534e00a2d50a3ad4b3b57893f242d
593c68582c2de4fff7c5dedcdb9dfb2acc541dfb
refs/heads/master
2021-02-05T17:26:09.958529
2020-02-19T06:23:01
2020-02-19T06:23:01
null
0
0
null
null
null
null
UTF-8
R
false
false
14,087
r
new.R
#New methods and functions #------Preliminary template---- weightit2XXX <- function(covs, treat...) { stop("method = \"XXX\" isn't ready to use yet.", call. = FALSE) } #------Template---- weightit2XXX <- function(covs, treat, s.weights, subset, estimand, focal, moments, int, ...) { A <- list(...) covs <- covs[subset, , drop = FALSE] treat <- factor(treat[subset]) covs <- cbind(covs, int.poly.f(covs, poly = moments, int = int)) covs <- apply(covs, 2, make.closer.to.1) if (any(vars.w.missing <- apply(covs, 2, function(x) anyNA(x)))) { missing.ind <- apply(covs[, vars.w.missing, drop = FALSE], 2, function(x) as.numeric(is.na(x))) covs[is.na(covs)] <- 0 covs <- cbind(covs, missing.ind) } new.data <- data.frame(treat, covs) new.formula <- formula(new.data) for (f in names(formals(PACKAGE::FUNCTION))) { if (is_null(A[[f]])) A[[f]] <- formals(PACKAGE::FUNCTION)[[f]] } A[names(A) %in% names(formals(weightit2optweight))] <- NULL A[["formula"]] <- new.formula A[["data"]] <- new.data A[["estimand"]] <- estimand A[["s.weights"]] <- s.weights[subset] A[["focal"]] <- focal A[["verbose"]] <- TRUE if (check.package("optweight")) { out <- do.call(PACKAGE::FUNCTION, A, quote = TRUE) obj <- list(w = out[["weights"]], fit.obj = out) return(obj) } } #------Under construction---- #Subgroup Balancing PS weightit2sbps <- function(covs, treat, s.weights, subset, estimand, focal, stabilize, ...) { A <- list(...) fit.obj <- NULL covs <- covs[subset, , drop = FALSE] t <- factor(treat[subset]) if (!is_binary(t)) stop("Subgroup balancing propensity score weighting is not yet compatible with non-binary treatments.", call. = FALSE) if (any(vars.w.missing <- apply(covs, 2, function(x) anyNA(x)))) { missing.ind <- apply(covs[, vars.w.missing, drop = FALSE], 2, function(x) as.numeric(is.na(x))) covs[is.na(covs)] <- 0 covs <- cbind(covs, missing.ind) } covs <- apply(covs, 2, make.closer.to.1) smd <- function(x, t, w, estimand, std = TRUE) { m <- vapply(levels(t), function(t.lev) w.m(x[t==t.lev], w = w[t==t.lev]), numeric(1L)) mdiff <- abs(diff(m)) if (check_if_zero(mdiff)) return(0) else { if (!std) sd <- 1 else if (estimand == "ATT") sd <- sd(x[t==1]) else if (estimand == "ATC") sd <- sd(x[t==0]) else sd <- sqrt(.5 * (var(x[t==1]) + var(x[t==0]))) return(mdiff/sd) } } loss <- A[["loss"]] loss <- match_arg(loss, c("weighting", "matching")) if (loss == "matching") { F_ <- function(covs, sub, t, w) { #Overall Balance of covs Mk <- apply(covs, 2, function(x) smd(x, t, w, estimand)) #Subgroup Balance Mkr <- unlist(lapply(levels(sub), function(s) {apply(covs, 2, function(x) smd(x[sub==s], t[sub==s], w[sub==s], estimand))})) return(sum(c(Mk, Mkr) ^ 2)) } } else if (loss == "weighting") { F_ <- function(covs, sub, t, w) { #Overall Balance of covs Mk <- apply(covs, 2, function(x) smd(x, t, w, estimand)) #Overall balance of subgroups Mr <- vapply(levels(sub), function(s) {smd(as.numeric(sub == s), t, w, std = FALSE)}, numeric(1L)) #Subgroup Balance Mkr <- unlist(lapply(levels(sub), function(s) {apply(covs, 2, function(x) smd(x[sub==s], t[sub==s], w[sub==s], estimand))})) return(sum(c(Mk, Mr, Mkr) ^ 2)) } } else stop() #Process subgroup subgroup <- process.by(by = A[["subgroup"]], data = covs, treat = t, by.arg = "subgroup")$by.factor overall.weights <- subgroup.weights <- NULL if (is_not_null(A[["overall.ps"]])) { if ((is.matrix(A[["overall.ps"]]) || is.data.frame(A[["overall.ps"]])) && ncol(A[["overall.ps"]]) == nlevels(t) && all(colnames(A[["overall.ps"]] %in% levels(t)))) { ps.mat <- A[["overall.ps"]] } else if (is.numeric(A[["overall.ps"]])) { ps.mat <- matrix(NA_real_, nrow = length(t), ncol = nlevels(t), dimnames = list(NULL, levels(t))) ps.mat[, 2] <- A[["overall.ps"]] ps.mat[, 1] <- 1 - A[["overall.ps"]] } else { stop() } overall.weights <- get_w_from_ps(ps.mat, t, estimand, focal) } if (is_not_null(A[["subgroup.ps"]])) { if ((is.matrix(A[["subgroup.ps"]]) || is.data.frame(A[["subgroup.ps"]])) && ncol(A[["subgroup.ps"]]) == nlevels(t) && all(colnames(A[["subgroup.ps"]] %in% levels(t)))) { ps.mat <- A[["subgroup.ps"]] } else if (is.numeric(A[["subgroup.ps"]])) { ps.mat <- matrix(NA_real_, nrow = length(t), ncol = nlevels(t), dimnames = list(NULL, levels(t))) ps.mat[, 2] <- A[["subgroup.ps"]] ps.mat[, 1] <- 1 - A[["subgroup.ps"]] } else { stop() } subgroup.weights <- get_w_from_ps(ps.mat, t, estimand, focal) } if (is_not_null(A[["overall.weights"]])) { if (!is.numeric(A[["overall.weights"]])) { stop() } overall.weights <- A[["overall.weights"]] } if (is_not_null(A[["subgroup.weights"]])) { if (!is.numeric(A[["subgroup.weights"]])) { stop() } subgroup.weights <- A[["subgroup.weights"]] } if (is_null(overall.weights) || is_null(subgroup.weights)) { #Process w.method w.method <- A[["w.method"]] check.acceptable.method(w.method, msm = FALSE, force = FALSE) if (is.character(w.method)) { w.method <- method.to.proper.method(w.method) attr(w.method, "name") <- w.method } else if (is.function(w.method)) { w.method.name <- paste(deparse(substitute(w.method))) check.user.method(w.method) attr(w.method, "name") <- w.method.name } if (loss == "matching") { t.bin <- binarize(t) overall.fit <- weightit.fit(covs = covs, treat = t, method = "ps", treat.type = "binary", s.weights = s.weights, by.factor = factor(rep(1, length(t))), estimand = estimand, focal = focal, stabilize = stabilize, ps = NULL, moments = 1, int = FALSE) overall.ps <- overall.fit$ps overall.match <- Matching::Match(Tr = t.bin, X = matrix(c(overall.ps, as.numeric(subgroup)), ncol = 2), estimand = estimand, caliper = .25, M = 1, replace = FALSE, exact = c(FALSE, TRUE), ties = TRUE) overall.weights <- cobalt::get.w(overall.match) subgroup.fit <- weightit.fit(covs = covs, treat = t, method = "ps", treat.type = "binary", s.weights = s.weights, by.factor = subgroup, estimand = estimand, focal = focal, stabilize = stabilize, ps = NULL, moments = 1, int = FALSE) subgroup.ps <- subgroup.fit$ps subgroup.match <- Matching::Match(Tr = t.bin, X = matrix(c(subgroup.ps, as.numeric(subgroup)), ncol = 2), estimand = estimand, caliper = .25, M = 1, replace = FALSE, exact = c(FALSE, TRUE), ties = TRUE) subgroup.weights <- cobalt::get.w(subgroup.match) } if (loss == "weighting") { #Estimate overall weights overall.fit <- weightit.fit(covs = covs, treat = t, method = w.method, treat.type = "binary", s.weights = s.weights, by.factor = factor(rep(1, length(t))), estimand = estimand, focal = focal, stabilize = stabilize, ps = NULL, moments = 1, int = FALSE) overall.weights <- overall.fit$w #Estimate subgroup weights subgroup.fit <- weightit.fit(covs = covs, treat = t, method = w.method, treat.type = "binary", s.weights = s.weights, by.factor = subgroup, estimand = estimand, focal = focal, stabilize = stabilize, ps = NULL, moments = 1, int = FALSE) subgroup.weights <- subgroup.fit$w } } #Find combinations that minimize loss n.subgroups <- nunique(subgroup) if (n.subgroups > 8) { #Stochastic search L1 <- 10 L2 <- 5 S_ <- setNames(rep("overall", nlevels(subgroup)), levels(subgroup)) rep <- 0 no.change.streak <- 0 current.loss <- Inf while (rep <= L1 && no.change.streak <= L2) { rep <- rep + 1 if (is_null(get0("last.loss"))) last.loss <- Inf else last.loss <- current.loss rand.subs <- sample(levels(subgroup)) S__ <- setNames(sample(c("overall", "subgroup"), length(S_), replace = TRUE), rand.subs) for (i in 1:length(S__)) { S__[i] <- "overall" to.overall <- subgroup %in% rand.subs[S__[rand.subs] == "overall"] w_ <- subgroup.weights w_[to.overall] <- overall.weights[to.overall] loss.o <- F_(covs, subgroup, t, w_) S__[i] <- "subgroup" to.overall <- subgroup %in% rand.subs[S__[rand.subs] == "overall"] w_ <- subgroup.weights w_[to.overall] <- overall.weights[to.overall] loss.s <- F_(covs, subgroup, t, w_) if (loss.o < loss.s) { S__[i] <- "overall" if (loss.o < current.loss) { current.loss <- loss.o attr(current.loss, "S") <- S__ } } else { S__[i] <- "subgroup" if (loss.s < current.loss) { current.loss <- loss.s attr(current.loss, "S") <- S__ } } } to.overall <- subgroup %in% rand.subs[S__[rand.subs] == "overall"] w_ <- subgroup.weights w_[to.overall] <- overall.weights[to.overall] current.loss <- F_(covs, subgroup, t, w_) if (check_if_zero(current.loss - last.loss)) no.change.streak <- no.change.streak + 1 print(current.loss) print(S__) } best.S <- attr(current.loss, "S") to.overall <- subgroup %in% rand.subs[best.S[rand.subs] == "overall"] w <- subgroup.weights w[to.overall] <- overall.weights[to.overall] } else { S <- setNames(do.call("expand.grid", lapply(integer(n.subgroups), function(x) (c("overall", "subgroup")))), levels(subgroup)) print(S) w.list <<- lapply(seq_len(nrow(S)), function(i) { to.overall <- subgroup %in% levels(subgroup)[S[i, levels(subgroup)] == "overall"] w_ <- subgroup.weights w_[to.overall] <- overall.weights[to.overall] return(w_) }) loss.val <- vapply(w.list, function(w_) F_(covs, subgroup, t, w_), numeric(1L)) best.loss <- which.min(loss.val) w <- w.list[[best.loss]] if (is_not_null(overall.fit$ps)) { to.overall <- subgroup %in% levels(subgroup)[S[best.loss, levels(subgroup)] == "overall"] p.score <- subgroup.fit$ps p.score[to.overall] <- overall.fit$ps[to.overall] } else p.score <- NULL } obj <- list(w = w , ps = p.score #, fit.obj = fit.obj ) return(obj) } #------Ready for use, but not ready for CRAN---- #KBAL weightit2kbal <- function(covs, treat, s.weights, subset, estimand, focal, ...) { A <- list(...) covs <- covs[subset, , drop = FALSE] treat <- factor(treat)[subset] covs <- apply(covs, 2, make.closer.to.1) if (any(vars.w.missing <- apply(covs, 2, function(x) anyNA(x)))) { missing.ind <- apply(covs[, vars.w.missing, drop = FALSE], 2, function(x) as.numeric(is.na(x))) covs[is.na(covs)] <- 0 covs <- cbind(covs, missing.ind) } if ("kbal.method" %in% names(A)) { names(A)[names(A) == "kbal.method"] <- "method" } for (f in names(formals(KBAL::kbal))) { if (is_null(A[[f]])) A[[f]] <- formals(KBAL::kbal)[[f]] } A[names(A) %nin% setdiff(names(formals(KBAL::kbal)), c("X", "D"))] <- NULL if (check.package("KBAL")) { if (hasName(A, "method")) { if (A[["method"]] == "el") check.package(c("glmc", "emplik")) } if (estimand == "ATT") { w <- rep(1, length(treat)) control.levels <- levels(treat)[levels(treat) != focal] fit.list <- setNames(vector("list", length(control.levels)), control.levels) covs[treat == focal,] <- covs[treat == focal, , drop = FALSE] * s.weights[subset][treat == focal] * sum(treat == focal)/sum(s.weights[subset][treat == focal]) for (i in control.levels) { treat.in.i.focal <- treat %in% c(focal, i) treat_ <- ifelse(treat[treat.in.i.focal] == i, 0L, 1L) covs_ <- covs[treat.in.i.focal, , drop = FALSE] colinear.covs.to.remove <- colnames(covs_)[colnames(covs_) %nin% colnames(make_full_rank(covs_[treat_ == 0, , drop = FALSE]))] covs_ <- covs_[, colnames(covs_) %nin% colinear.covs.to.remove, drop = FALSE] kbal.out <- do.call(KBAL::kbal, c(list(X = covs_, D = treat_), args)) w[treat == i] <- (kbal.out$w / s.weights[subset])[treat_ == 0L] fit.list[[i]] <- kbal.out } } else if (estimand == "ATE") { w <- rep(1, length(treat)) fit.list <- setNames(vector("list", nlevels(treat)), levels(treat)) for (i in levels(treat)) { covs_i <- rbind(covs, covs[treat==i, , drop = FALSE]) treat_i <- c(rep(1, nrow(covs)), rep(0, sum(treat==i))) colinear.covs.to.remove <- colnames(covs_i)[colnames(covs_i) %nin% colnames(make_full_rank(covs_i[treat_i == 0, , drop = FALSE]))] covs_i <- covs_i[, colnames(covs_i) %nin% colinear.covs.to.remove, drop = FALSE] covs_i[treat_i == 1,] <- covs_i[treat_i == 1,] * s.weights[subset] * sum(treat_i == 1) / sum(s.weights[subset]) kbal.out_i <- do.call(KBAL::kbal, c(list(X = covs_i, D = treat_i), args)) w[treat == i] <- kbal.out_i$w[treat_i == 0] / s.weights[subset][treat == i] fit.list[[i]] <- kbal.out_i } } } obj <- list(w = w) return(obj) }
8b6038c1eb0e95bba11102f7842bc9a3f37e3528
ffdea92d4315e4363dd4ae673a1a6adf82a761b5
/data/genthat_extracted_code/DiagrammeR/examples/trav_both_edge.Rd.R
df9aea57ce9b84ec06148ea819c972ed1b859911
[]
no_license
surayaaramli/typeRrh
d257ac8905c49123f4ccd4e377ee3dfc84d1636c
66e6996f31961bc8b9aafe1a6a6098327b66bf71
refs/heads/master
2023-05-05T04:05:31.617869
2019-04-25T22:10:06
2019-04-25T22:10:06
null
0
0
null
null
null
null
UTF-8
R
false
false
3,716
r
trav_both_edge.Rd.R
library(DiagrammeR) ### Name: trav_both_edge ### Title: Traverse from one or more selected nodes onto adjacent edges ### Aliases: trav_both_edge ### ** Examples # Set a seed set.seed(23) # Create a simple graph graph <- create_graph() %>% add_n_nodes( n = 2, type = "a", label = c("asd", "iekd")) %>% add_n_nodes( n = 3, type = "b", label = c("idj", "edl", "ohd")) %>% add_edges_w_string( edges = "1->2 1->3 2->4 2->5 3->5", rel = c(NA, "A", "B", "C", "D")) # Create a data frame with node ID values # representing the graph edges (with `from` # and `to` columns), and, a set of numeric values df <- data.frame( from = c(1, 1, 2, 2, 3), to = c(2, 3, 4, 5, 5), values = round(rnorm(5, 5), 2)) # Join the data frame to the graph's internal # edge data frame (edf) graph <- graph %>% join_edge_attrs(df = df) # Show the graph's internal edge data frame graph %>% get_edge_df() # Perform a simple traversal from nodes to # adjacent edges with no conditions on the # nodes traversed to graph %>% select_nodes_by_id(nodes = 3) %>% trav_both_edge() %>% get_selection() # Traverse from node `2` to any adjacent # edges, filtering to those edges that have # NA values for the `rel` edge attribute graph %>% select_nodes_by_id(nodes = 2) %>% trav_both_edge( conditions = is.na(rel)) %>% get_selection() # Traverse from node `2` to any adjacent # edges, filtering to those edges that have # numeric values greater than `6.5` for # the `rel` edge attribute graph %>% select_nodes_by_id(nodes = 2) %>% trav_both_edge( conditions = values > 6.5) %>% get_selection() # Traverse from node `5` to any adjacent # edges, filtering to those edges that # have values equal to `C` for the `rel` # edge attribute graph %>% select_nodes_by_id(nodes = 5) %>% trav_both_edge( conditions = rel == "C") %>% get_selection() # Traverse from node `2` to any adjacent # edges, filtering to those edges that # have values in the set `B` and `C` for # the `rel` edge attribute graph %>% select_nodes_by_id(nodes = 2) %>% trav_both_edge( conditions = rel %in% c("B", "C")) %>% get_selection() # Traverse from node `2` to any adjacent # edges, and use multiple conditions for the # traversal graph %>% select_nodes_by_id(nodes = 2) %>% trav_both_edge( conditions = rel %in% c("B", "C") & values > 4.0) %>% get_selection() # Traverse from node `2` to any adjacent # edges, and use multiple conditions with # a single-length vector graph %>% select_nodes_by_id(nodes = 2) %>% trav_both_edge( conditions = rel %in% c("B", "C") | values > 4.0) %>% get_selection() # Traverse from node `2` to any adjacent # edges, and use a regular expression as # a filtering condition graph %>% select_nodes_by_id(nodes = 2) %>% trav_both_edge( conditions = grepl("B|C", rel)) %>% get_selection() # Create another simple graph to demonstrate # copying of node attribute values to traversed # edges graph <- create_graph() %>% add_path(n = 4) %>% select_nodes_by_id(nodes = 2:3) %>% set_node_attrs_ws( node_attr = value, value = 5) # Show the graph's internal edge data frame graph %>% get_edge_df() # Show the graph's internal node data frame graph %>% get_node_df() # Perform a traversal from the nodes to # the adjacent edges while also applying # the node attribute `value` to the edges (in # this case summing the `value` of 5 from # all contributing nodes adding as an edge # attribute) graph <- graph %>% trav_both_edge( copy_attrs_from = value, agg = "sum") # Show the graph's internal edge data frame # after this change graph %>% get_edge_df()
3c821f4119fcef35e12ce910df26531bfe5c7a67
33464d95ff20ff7feda3ad658fb3f2de5b075308
/code/raw/good_turing.R
bba6200e8fd54a76e20cdadab86e5752ee3abf6b
[]
no_license
YevgenyY/DS_Capstone
514cd599fa61450f7eec84e89e293de12b229376
883fbd590dc61135e5c9d2417460a8eef10d3b39
refs/heads/master
2021-01-10T14:41:52.539675
2016-04-22T10:54:11
2016-04-22T10:54:11
53,430,751
1
0
null
null
null
null
UTF-8
R
false
false
890
r
good_turing.R
library(quanteda) library(parallel) library(foreach) setwd("~/Coursera/DS_Capstone/") source("code/raw/stats_helper.R") #load(file="data/freq.Rda") load(file="data/f12345.Rda") # use parallel computation tmp <- names(f2) no_cores <- detectCores()-1 cl <- makeCluster(no_cores) ########### Good-Turing smoothing ########## # Find Nc for 1-grams clusterExport(cl, "f1") f1Nc <- parSapply(cl, f1, function(x) { t <- f1 == x return(sum(t)) }) # find discount for 1-grams clusterExport(cl, "f1Nc") f1Dc <- parSapply(cl, names(f1), function(x) { Nc <- f1Nc[x] # "said" has c=76778, but Nc=1 cp1<- f1[x] + 1 # c + 1 #i <- match(cp1, f1) for(i in 1:length(f1)) { if(cp1 <= f1[i]) break } if (i == length(f1)) Ncp1 <- 0 else { j <- names(f1[i]) Ncp1 <- f1Nc[j] } return(Ncp1/Nc) }) pf1 <- (f1+1)*f1Dc save(pf1, file="data/good_turing.pf1.Rda")
03f7751647d1a3bb791a8576fa8a55c201787bfb
839db0fc473db0c3b9cb169a091bbb6046213ff2
/validation-scripts.previous/bench-merge-above5500/Rscripts/cmp_extension_assembly.R
2a9fa77c9d4fdfceff39314b62ea07d35ad64e16
[]
no_license
Transipedia/KaMRaT
3a5754a14119cda0baf9fdd13f2aa6b5f735583f
6da06e353d9dd861285ac057ba19e790463cf4ed
refs/heads/master
2023-06-10T09:09:59.111236
2023-05-22T18:47:16
2023-05-22T18:47:16
241,427,846
6
1
null
null
null
null
UTF-8
R
false
false
5,238
r
cmp_extension_assembly.R
rm(list = ls()) library(Biostrings) library(tidyr) library(ggplot2) library(patchwork) # work.dir <- "/data/work/I2BC/haoliang.xue/kamrat-new-res/Results/bench-merge-above5500/" work.dir <- "/store/plateformes/CALCUL/SSFA_KaMRaT/Results/bench-merge-above5500/" # work.dir <- "../../../../Ariticles/KaMRaT/RevisedAnalysis/bench-merge/" min_len <- 61 evaluate_perfalign <- function(ctg.fa, align.res) { perfect.align <- (align.res$qlen == align.res$qend & align.res$qstart == 1 & align.res$qlen == align.res$length & align.res$qlen == align.res$nident & align.res$pident == 100 & align.res$qlen > min_len) return(sum(perfect.align) / sum(width(ctg.fa) > min_len) * 100) } evaluate_compalign <- function(ctg.fa, align.res) { return(sum(align.res$qcovs == 100 & align.res$qlen > min_len) / sum(width(ctg.fa) > min_len) * 100) } abd <- 1 stats.res <- NULL for (dpt in c(0.05, 0.2, 0.4, 0.6, 0.8, 1)) { # KaMRaT none cat(dpt, "KaMRaT none", ":") ctg.path <- paste0(work.dir, "kamrat_res_err-free_1-", abd, "/depth_", dpt, "/ctg-seq.none.fa") align.path <- paste0(work.dir, "kamrat_res_err-free_1-", abd, "/depth_", dpt, "/ctg-aligned.none.tsv") ctg.fa <- readDNAStringSet(ctg.path) align.res <- read.table(align.path, header = TRUE, row.names = 1) cat("\t", length(ctg.fa), nrow(align.res), "\n") stats.res <- rbind(stats.res, data.frame("depth" = dpt, "mode" = "KaMRaT none", "nb.ctg" = length(ctg.fa), "ctg.median.len" = median(nchar(ctg.fa)), "perf.align" = evaluate_perfalign(ctg.fa, align.res), "comp.align" = evaluate_compalign(ctg.fa, align.res))) # KaMRaT with intervention for (mode in c("pearson", "spearman", "mac")) { cat(dpt, "KaMRaT", paste(mode, 0.2, sep = ":"), ":") ctg.path <- paste0(work.dir, "kamrat_res_err-free_1-", abd, "/depth_", dpt, "/ctg-seq.", mode, "_0.2", ".fa") align.path <- paste0(work.dir, "kamrat_res_err-free_1-", abd, "/depth_", dpt, "/ctg-aligned.", mode, "_0.2", ".tsv") ctg.fa <- readDNAStringSet(ctg.path) align.res <- read.table(align.path, header = TRUE, row.names = 1) cat("\t", length(ctg.fa), nrow(align.res), "\n") stats.res <- rbind(stats.res, data.frame("depth" = dpt, "mode" = paste0("KaMRaT ", mode, ":0.2"), "nb.ctg" = length(ctg.fa), "ctg.median.len" = median(nchar(ctg.fa)), "perf.align" = evaluate_perfalign(ctg.fa, align.res), "comp.align" = evaluate_compalign(ctg.fa, align.res))) } # rnaSPAdes for (mode in c("allreads", paste0("allkmers-1-", abd))) { cat(dpt, "SPAdes", mode, ":") ctg.path <- paste0(work.dir, "spades_res/err-free/depth_", dpt, "/", mode, "/transcripts.fasta") align.path <- paste0(work.dir, "spades_res/err-free/depth_", dpt, "/", mode, "/blastn_align.tsv") ctg.fa <- readDNAStringSet(ctg.path) align.res <- read.table(align.path, header = TRUE, row.names = 1) cat("\t", length(ctg.fa), nrow(align.res), "\n") stats.res <- rbind(stats.res, data.frame("depth" = dpt, "mode" = paste0("rnaSPAdes ", strsplit(mode, split = "-")[[1]][1]), "nb.ctg" = length(ctg.fa), "ctg.median.len" = median(nchar(ctg.fa)), "perf.align" = evaluate_perfalign(ctg.fa, align.res), "comp.align" = evaluate_compalign(ctg.fa, align.res))) } } write.csv(stats.res, paste0(work.dir, "results/2_newcmp_with_without_intervention_1-", abd, ".csv"), quote = FALSE) pdf(paste0(work.dir, "results/2_newcmp_with_without_intervention.pdf"), width=9, height=7) plt1 <- ggplot(data = stats.res, aes(x = depth, y = perf.align, color = mode)) + geom_line(linewidth = 1) + geom_point() + scale_x_continuous(breaks = c(0.05, 0.2, 0.4, 0.6, 0.8, 1)) + scale_color_manual(values = c("KaMRaT none" = "#e66101", "KaMRaT mac:0.2" = "#fdb863", "KaMRaT pearson:0.2" = "#b2abd2", "KaMRaT spearman:0.2" = "#5e3c99", "rnaSPAdes allkmers" = "#808080", "rnaSPAdes allreads" = "#000000")) + ylim(c(70, 100)) + ylab("%perfect alignment") + theme_light() + theme(text = element_text(size = 15, family = "sans"), panel.grid.major = element_blank(), panel.grid.minor = element_blank()) plt2 <- ggplot(data = stats.res, aes(x = depth, y = comp.align, color = mode)) + geom_line(linewidth = 1) + geom_point() + scale_x_continuous(breaks = c(0.05, 0.2, 0.4, 0.6, 0.8, 1)) + scale_color_manual(values = c("KaMRaT none" = "#e66101", "KaMRaT mac:0.2" = "#fdb863", "KaMRaT pearson:0.2" = "#b2abd2", "KaMRaT spearman:0.2" = "#5e3c99", "rnaSPAdes allkmers" = "#808080", "rnaSPAdes allreads" = "#000000")) + ylab("%complete alignment") + theme_light() + theme(text = element_text(size = 15, family = "sans"), panel.grid.major = element_blank(), panel.grid.minor = element_blank()) plot(plt1/plt2) dev.off()
e26717063d1fccd5c984a6c05d306f20a5c8a2f6
4343c3a44a0eb0effaeb45a4f695a1b42386860d
/man/searchData.Rd
dbb6328390d3da7625a1e52b81aea0e58adc3f23
[ "Apache-2.0" ]
permissive
abossenbroek/GeneralTree
1b19a86cc23bf713e3352311e69d3edbcf9fbf79
3efe2d9cc5a9082cba6995961b19159bfc9224a9
refs/heads/master
2020-12-25T16:57:33.837972
2016-09-09T12:27:45
2016-09-09T12:27:45
58,137,112
3
1
null
2016-09-09T16:05:57
2016-05-05T14:32:37
R
UTF-8
R
false
true
499
rd
searchData.Rd
% Generated by roxygen2: do not edit by hand % Please edit documentation in R/GeneralTree.R \name{searchData} \alias{searchData} \title{Search for an id in starting at a point in the tree and return the data matching the id.} \usage{ searchData(self, id) } \arguments{ \item{self}{the node where to start searching.} \item{id}{the id to look for.} } \value{ The data associated with an id. } \description{ Search for an id in starting at a point in the tree and return the data matching the id. }
693418832238b7075b2d29f72e4aa8f07c1df438
727e274d8dfbfc6eda1166844e185052133a8c32
/R/PublicationPlots/promoterCapSeq_parallelPlots.R
f142a3de144e42d3bc0ce7e06d44b813bd14ee69
[]
no_license
JCSMR-Tremethick-Lab/MCF10APromoterAnalysis
871c6162fbd42f114c75f386bf4b15d6d89fea88
218f8a241b5645796b2a4d8a6bb33caf38a146c1
refs/heads/master
2021-03-30T17:45:18.443603
2021-01-29T09:44:25
2021-01-29T09:44:25
56,828,470
0
0
null
2020-05-20T06:53:47
2016-04-22T05:26:30
R
UTF-8
R
false
false
1,971
r
promoterCapSeq_parallelPlots.R
library(data.table) library(ggparallel) # Figure 1 ---------------------------------------------------------------- setwd("/home/sebastian/Data/Collaborations/FSU/PromoterSeqCap/sortingTSVS for Tremethick paper /Figure 1/") sourceFiles <- list.files(".", pattern = ".tsv") dataList <- lapply(sourceFiles, function(x){ tab <- data.table::fread(x) return(tab) }) names(dataList) <- unlist(lapply(strsplit(sourceFiles, "\\."), function(x) x[1])) d1 <- do.call("data.table", dataList) d1 <- subset(d1, select = c(1, grep("group1", colnames(d1)))) colNames <- colnames(d1)[2:8] colNames <- unlist(lapply(strsplit(colNames, "_"), function(x) paste(x[1:2], collapse = "_"))) colnames(d1) <- c("gene", colNames) ggparallel(list("A_Inp", "A_H2AZ", "CA1a_Inp", "CA1a_H2AZ", "shH2AZ_Inp", "TGFb_Inp", "TGFb_H2AZ"), data = d1) ggparallel(list("A_Inp", "CA1a_Inp", "shH2AZ_Inp", "TGFb_Inp"), data = d1) ggparallel(list("A_H2AZ", "CA1a_H2AZ", "TGFb_H2AZ"), data = d1) ggparallel(list("A_H2AZ", "CA1a_H2AZ", "TGFb_H2AZ"), data = d1, method = "hammock", ratio = 0.1) ggparallel(list("A_Inp", "CA1a_Inp", "shH2AZ_Inp", "TGFb_Inp"), data = d1, method = "hammock", ratio = 0.1) # attempt at using plot.ly for interactive vis library(plotly) p <- m1 %>% plot_ly(width = 1920, height = 1080) %>% add_trace(type = 'parcoords', line = list(showscale = TRUE, reversescale = TRUE, color = ~order.x, colorscale = 'Jet', cmin = 0, cmax = 20000), dimensions = list( list(tickvals = c(1:7), label = "group1", values = ~group1.x), list(tickvals = c(1:7), label = "group2", values = ~group1.y), list(tickvals = c(1:7), label = "group3", values = ~group1) ) ) as_widget(p)
b5fb203b27fd53f74394bc02b8ac037eaa83b41a
03715ef672cc49b9de1e46bab7ebe68eb69b90ef
/man/rawtocomposite.Rd
616d71f7f9938472e0f1f8da34eba1f8fff09752
[]
no_license
cran/waveformlidar
501b867be68524862ec10fb12a4aefb9837b19ec
7a0fee0455142f1ade951ccd191a2299f2d8ed1e
refs/heads/master
2021-07-11T16:10:17.124894
2020-08-01T08:20:03
2020-08-01T08:20:03
184,310,247
2
0
null
null
null
null
UTF-8
R
false
true
2,494
rd
rawtocomposite.Rd
% Generated by roxygen2: do not edit by hand % Please edit documentation in R/rawtocomposite.R \name{rawtocomposite} \alias{rawtocomposite} \title{rawtocomposite} \usage{ rawtocomposite(voxr, inten_index = 2) } \arguments{ \item{voxr}{the object from the waveformvoxel.} \item{inten_index}{the value (1,2,3,4,...) to represnt the intensity of composite waveforms.It is a integer from 1 to 4 and default is 2. 1: the number of intensity of the voxel (generally is not useful); 2: the maximum intensity of the waveform voxel; 3: the mean intensity of the waveform voxel; 4: the total intensity of voxel(the last one is also not prefered in most cases)} } \value{ A dataframe with first three columns including geolocation xyz of the first Non-NA intensity (Highest position) and intensities along the height bins, other non-NA values are intensities for the rest columns. \item{x}{The x position of the first Non-NA intensity or highest intensity position in one waveform} \item{y}{The y position of the first Non-NA intensity or highest intensity position in one waveform} \item{z}{The z position of the first Non-NA intensity or highest intensity position in one waveform} \item{intensity 1}{The intnesity of first height bin} \item{intensity 2}{The intensity of second height bin} \item{...}{Intensities along the height bin} } \description{ The function allows you to convert point cloud after waveformvoxel or raw waveforms into composite waveforms (with vertical distribution of intensity) by reducing the effect of off-naid angle of emitted laser. The conversion is based on the waveform voxellization product. Four kinds of values you can chose to represent the intensity of composite waveform: the number of intensity (generally is not useful), mean intensity, maximum intensity and total intensity (the last one is also not prefered in most of cases). } \examples{ data(return) ###import raw return waveforms data(geo) ###import corresponding reference geolocation colnames(geo)[2:9]<-c("x","y","z","dx","dy","dz","or","fr") ### you should know which columns corresponding to above column names before ### run the hyperpointcloud when you used your own new datasets. hpr<-hyperpointcloud(waveform=return,geo=geo) ##beofre run waveformvoxel, we need to create hyperpointcloud first ##this exampel we just used 100000 points to reduce processing time voxr<-waveformvoxel(hpc=hpr,res=c(1,1,0.3)) rtc<-rawtocomposite(voxr) }
9a7d433030fbf6dfb7d87b074cb8ee50dc48ad37
514e2fc6b95ccaa013c4000ed7001002e42c72b8
/man/plot.landscape.Rd
ceb60b49d2739c6356dc629d3d2df8f8f52aced2
[ "MIT" ]
permissive
fdschneider/caspr
7493a32dbd8df8977135c4f57706e2b4bcc5bbec
e191c7cfaebe1499801e08d5fda1728d4697051f
refs/heads/master
2021-01-15T14:50:33.742811
2016-01-13T16:13:20
2016-01-13T16:13:20
37,334,564
4
1
null
2015-06-26T16:13:08
2015-06-12T17:13:56
R
UTF-8
R
false
true
1,217
rd
plot.landscape.Rd
% Generated by roxygen2: do not edit by hand % Please edit documentation in R/landscape.R \name{plot.landscape} \alias{plot.landscape} \title{Plotting an objects of class "landscape"} \usage{ \method{plot}{landscape}(x, cols = "auto", grid = FALSE, axis = FALSE, add = FALSE, ani = FALSE, ...) } \arguments{ \item{x}{A landscape object.} \item{cols}{A color vector. If \code{"auto"}, then a grayscale vector will be used.} \item{grid}{If TRUE, plot a grid over the cells. Defaults to FALSE.} \item{axis}{If TRUE, plot x and y axis with coordinates. Defaults to FALSE.} \item{add}{If TRUE, no primary plot will be called. The change in cells will be plotted over an existing plot. Used for animated plotting in the screen plotting device.} \item{ani}{If TRUE, an adjustment constant is added when producing a pixel accurate png or gif file. Required when the function is used to plot animated figures.} } \value{ A landscape object of dimensions \code{width} x \code{height} with random distribution of \code{states}, in the relative ratio given in \code{cover}. } \description{ Plotting an objects of class "landscape" } \examples{ obj <- init_landscape(c("+","0","-"), c(0.5,0.25,0.25)) plot(obj) }
17b2140542e8e7101d458a38a6bb27c1774da963
458e881217581e0b823ad6661ea4f1aa53a34cbf
/cachematrix.R
aadf36c32ea818fa504f2dc8f5f8d16f3479e495
[]
no_license
course-student/ProgrammingAssignment2
4cb833e777a335ac757a2df0f2171fa6e8661007
03086008e8e81694f55b37ee374f1ca7936cf563
refs/heads/master
2020-12-25T21:34:07.403889
2015-07-20T20:26:41
2015-07-20T20:26:41
39,405,555
0
0
null
2015-07-20T19:59:19
2015-07-20T19:59:19
null
UTF-8
R
false
false
1,040
r
cachematrix.R
## These functions calculates the inverse of a matrix and caches it ## makeCacheMatrix creates a special "matrix" that contains the inverse of the original matrix cached makeCacheMatrix <- function(x = matrix()) { inverse <- NULL set <- function(y){ x <<- y inverse <<- NULL } get <- function() x setinverse <- function(s) inverse <<- s getinverse <- function() inverse list(set = set, get = get, setinverse = setinverse, getinverse = getinverse) } ## cacheSolve calculates de inverse of the matrix, retorning the cached value if it exists ## assuming that the matrix supplied is always invertible cacheSolve <- function(x, ...) { ## Return a matrix that is the inverse of 'x' inverse <- x$getinverse() if(!is.null(inverse)){ message("getting cached data") return(inverse) } data <- x$get() inverse <- solve(data, ...) x$setinverse(inverse) inverse }