thewall commited on
Commit
680f288
·
1 Parent(s): fc32727

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +71 -0
README.md CHANGED
@@ -1,3 +1,74 @@
1
  ---
2
  license: openrail
3
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  license: openrail
3
  ---
4
+ DEEPBIND v0.11
5
+ --------------
6
+
7
+ The deepbind command-line executable can be used to score DNA/RNA sequences
8
+ according to any RBP/TF model listed in the DeepBind web repository:
9
+
10
+ http://tools.genes.toronto.edu/deepbind
11
+
12
+ For each input sequence, the deepbind executable scores each subsequence
13
+ of a pre-determined length (e.g. 20) and returns only the maximum or the
14
+ average over these per-position scores.
15
+
16
+ Larger scores indicated stronger binding. The scores themselves are on an
17
+ arbitrary scale, and vary from model to model due to variation in the
18
+ quality of training data for different proteins.
19
+
20
+
21
+ EXAMPLE
22
+ -------
23
+
24
+ To generate predictions with DeepBind, you need two things:
25
+
26
+ 1) a list of model IDs, and
27
+ 2) a list of DNA/RNA sequences.
28
+
29
+ The file example.ids contains 4 example model IDs, one
30
+ on each line, reproduced here:
31
+
32
+ D00210.001 # RBFOX1 (RNAcompete)
33
+ D00120.001 # MBNL1 (RNAcompete)
34
+ D00410.003 # GATA3 (SELEX)
35
+ D00328.003 # CTCF (SELEX)
36
+
37
+ The file example.seq contains 4 example sequences, which
38
+ were chosen such that the nth sequence scores highly for
39
+ the nth model. The file example.seq is reproduced here:
40
+
41
+ AGGUAAUAAUUUGCAUGAAAUAACUUGGAGAGGAUAGC
42
+ AGACAGAGCUUCCAUCAGCGCUAGCAGCAGAGACCAUU
43
+ GAGGTTACGCGGCAAGATAA
44
+ TACCACTAGGGGGCGCCACC
45
+
46
+ To generate 16 predictions (4 models, 4 sequences), run
47
+ the deepbind executable as follows:
48
+
49
+ % deepbind example.ids < example.seq
50
+ D00210.001 D00120.001 D00410.003 D00328.003
51
+ 7.451420 -0.166146 -0.408751 -0.026180
52
+ -0.155398 4.113817 0.516956 -0.248167
53
+ -0.140683 0.181295 5.885349 -0.026180
54
+ -0.174985 -0.152521 -0.379695 17.682623
55
+
56
+ To see details of each ID, use the --dump-info flag:
57
+
58
+ % deepbind --dump-info example.ids
59
+ ID Protein Type Species Family Class Experiment ...
60
+ D00210.001 RBFOX1 RBP Homo sapiens RRM RNAcompete ...
61
+ D00120.001 MBNL1 RBP Homo sapiens Znf RNAcompete ...
62
+ D00410.003 GATA3 TF Homo sapiens GATA SELEX ...
63
+ D00328.003 CTCF TF Homo sapiens C2H2 ZF SELEX ...
64
+
65
+
66
+
67
+ CHANGES v0.1 -> v0.11
68
+ ---------------------
69
+
70
+ - Fixed bug where last position in input sequence was
71
+ not evaluated for a score; suggested by Irene Kaplow.
72
+
73
+ - Added --window-size and --average flags based on feedback.
74
+