accession stringclasses 2
values | feature stringclasses 6
values | unit stringclasses 2
values | n_tumor int64 18 79 | n_normal int64 7 18 | mean_difference float64 -16.73 1.89 | ci_low float64 -19.91 1 | ci_high float64 -13.55 3.2 | p float64 0 0.2 | rank_p float64 0 0.24 | tumor_higher_pairs float64 0 18 ⌀ | leave_one_out_min float64 -17.31 1.44 | leave_one_out_max float64 -16.08 2.13 | q float64 0 0.2 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
GSE149507 | ANXA2 | published log2 expression | 18 | 18 | -0.89419 | -1.226697 | -0.561683 | 0.000027 | 0.000008 | 0 | -0.943858 | -0.787598 | 0.000047 |
GSE149507 | CIP2A | published log2 expression | 18 | 18 | 1.362926 | 0.997827 | 1.728026 | 0 | 0.000008 | 18 | 1.28622 | 1.436123 | 0.000003 |
GSE149507 | BIRC2 | published log2 expression | 18 | 18 | -0.24543 | -0.538224 | 0.047365 | 0.094914 | 0.081429 | 5 | -0.317507 | -0.172926 | 0.094914 |
GSE149507 | BIRC3 | published log2 expression | 18 | 18 | -1.052255 | -1.644807 | -0.459704 | 0.001607 | 0.001579 | 3 | -1.160472 | -0.886744 | 0.001928 |
GSE149507 | HALLMARK_TNFA_SIGNALING_VIA_NFKB | rank percentile points | 18 | 18 | -8.663136 | -11.920304 | -5.405968 | 0.000031 | 0.000053 | 1 | -9.403383 | -8.080313 | 0.000047 |
GSE149507 | GOBP_MACROAUTOPHAGY | rank percentile points | 18 | 18 | -0.998453 | -1.313854 | -0.683052 | 0.000004 | 0.000008 | 0 | -1.053869 | -0.941127 | 0.000012 |
GSE60052 | ANXA2 | published log2 expression | 79 | 7 | -1.791774 | -2.545847 | -1.037701 | 0.000557 | 0.000098 | null | -2.082284 | -1.679448 | 0.001114 |
GSE60052 | CIP2A | published log2 expression | 79 | 7 | 1.891986 | 0.580674 | 3.203299 | 0.011278 | 0.000646 | null | 1.444488 | 2.126667 | 0.016917 |
GSE60052 | BIRC2 | published log2 expression | 79 | 7 | -0.424059 | -1.127821 | 0.279703 | 0.201936 | 0.2356 | null | -0.595664 | -0.279925 | 0.201936 |
GSE60052 | BIRC3 | published log2 expression | 79 | 7 | -2.227489 | -2.874712 | -1.580266 | 0.000001 | 0.000725 | null | -2.365007 | -2.094659 | 0.000003 |
GSE60052 | HALLMARK_TNFA_SIGNALING_VIA_NFKB | rank percentile points | 79 | 7 | -16.728384 | -19.90849 | -13.548278 | 0 | 0 | null | -17.307983 | -16.079499 | 0.000001 |
GSE60052 | GOBP_MACROAUTOPHAGY | rank percentile points | 79 | 7 | -1.422107 | -3.794683 | 0.95047 | 0.200099 | 0.03507 | null | -2.346264 | -1.091594 | 0.201936 |
CIP2A transcript abundance and genetic fitness in small-cell lung cancer
This repository accompanies the manuscript CIP2A transcript elevation and model-specific genetic fitness effects in small-cell lung cancer . It contains derived analysis tables, figure inputs and outputs, executable scripts, and the public-source inventory. The Viewer exposes three analysis tables through the configuration selector. The analysis split identifies an evidence table.
Findings documented here
| Analysis | Reported result |
|---|---|
| Bulk, GSE149507 | Higher tumor CIP2A in 18/18 paired comparisons. Mean tumor minus adjacent lung +1.363 log₂ units, 95% CI +0.998 to +1.728. |
| Bulk, GSE60052 | 79 tumors and seven normal lung specimens. Mean difference +1.892 log₂ units, 95% CI +0.581 to +3.203. |
| Genetic screens | NCI-H211 and DMS 53 ranked first and second by negative Chronos effect among 23 SCLC models, and first and third by negative DEMETER2 effect among 11 matched models. The matched-panel Spearman correlation was 0.273. |
| Single-cell | Tumor-group CIP2A CPM exceeded epithelial-group CPM in four evaluable cases: two of four primary lung cases from the Chan atlas and two nonoverlapping GSE281740 cases with computational groups. Low-cell-cycle-marker results differed between the two GSE281740 cases. |
The model nominations were retrospective. The two GSE281740 cases used RNA-based arm scores and lineage markers for cell grouping. The patient was the unit of the single-cell comparisons. Repeated classifications and depth-matching draws quantify within-patient sensitivity. The evidence addresses transcript abundance and genetic fitness.
Reproduce the main numbers and figures
Run from the repository root. Use Python 3.12 and R 4.5.2 (base R).
python -m pip install -r requirements.txt
python scripts/check_results.py
Rscript scripts/plot_figures.R
check_results.py recomputes the bulk tests and adjusted P values, the CRISPR histology comparison and cross-screen ranks, and the patient-level single-cell measures and sensitivity summaries. It writes build/key_results.json. The R command writes 12 figures in PDF and PNG to build/figures/. The reference outputs are key_results.json and the figure directory. See figure and table mapping, upstream rerun instructions and local reproduction checks.
Data origins and citation
- Tissue cohorts from NCBI GEO. GSE149507 · GSE60052
- Genetic screens and model labels. DepMap Public 24Q2 · DEMETER2 · Bioconductor depmap · Cellosaurus
- Single-cell datasets and genome annotation. Chan et al. atlas · Kawasaki et al., GSE281740 · UCSC GRCh38
Cite the companion manuscript, the exact Hub revision used, and the relevant source studies. Obtain original matrices and screening resources from their upstream repositories under their source terms. Figure 3 plotting data contain published UMAP coordinates, public case IDs, group labels and binary CIP2A detection. See input provenance, additional single-cell and RNAi inputs and release hashes.
- Downloads last month
- 63