There is a newer version of the record available.

Published July 23, 2026 | Version v1.0

tjogzt/paralog-sl-predictor: v1.0 — Initial release

Authors/Creators

  • 1. Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology

Description

Delta Dependency (DD) — a single-subtraction metric for prioritizing paralog-based synthetic lethality candidates from DepMap CRISPR data across 23 solid tumor types.

Highlights: DD AUROC = 0.794, outperforming 8 published deep learning methods in CV3 benchmark CPTAC proteomics validation across 7 cohorts (n = 672) PRISM drug sensitivity: 553 compounds with selective activity in driver-mutant lines Therapeutic window analysis nominates ARID1A→ARID1B (TI = 4.13) Companion R package: devtools::install_github("tjogzt/paralogSL")

Data: DepMap 26Q1 · CPTAC · PRISM · TCGA | License: MIT

Files

tjogzt/paralog-sl-predictor-v1.0.zip

Files (4.1 MB)

Name Size
md5:27071aaf36dbd8955f9ce25ad530e66b
4.1 MB Preview Download

Additional details

Related works