depmap

Falsification-first research on cancer genetic vulnerabilities that replicate across independent CRISPR screens.

Package to predict dependencies between cell lines and genes using Network Representation Learning (NRL) based Link Prediction

multi-omic CTS extension and dependency-prediction ML model

DeepVul: A Multi-Task Transformer Model for Joint Prediction of Gene Essentiality and Drug Response

Multi-omics kinase target prioritization pipeline for triple-negative breast cancer (TNBC) — CTS scoring across 90 real RTK/NRTK kinases.

Multi-omic computational pipeline for prioritizing combination-therapy hypotheses in triple-negative breast cancer — kinase target scoring (CTS), regimen ranking (MDCOE/HCOS), DepMap/CPTAC validation, agentic literature discovery, and an in-development GNN-based drug-synergy predictor.

Evidence conflict audit for oncology drug target selection · 靶点尽调：七维公开证据、七条冲突规则、跨癌种比较，Python 标准库实现，零运行时依赖

Autonomous AI scientist hunting the gene dependencies childhood cancers can't live without. Every prediction is hash-locked before testing and published forever failures included.

Pharmacogenomic drug response modeling: GDSC2 × DepMap, elastic net vs XGBoost, tissue-adjusted biomarker scan, SHAP

Contradiction-aware cancer gene hypothesis triage app using public-data evidence layers.

Predicts CRISPR gene dependencies across 1,066 DepMap cancer models and ranks selective therapeutic hypotheses. Precision@10 0.965 on held-out models, validated against an independent Sanger screen.

Interactive web viewer for TF ChIP-seq binding programs at canonical protein-coding TSSs (Ensembl GRCh38.114). Streamlit + Plotly + DuckDB; ~1,300 TFs × ~19,700 TSSs; NMF programs/archetypes, GTEx + DepMap overlays, atlas-wide TF×TF network.

Stability and counterfactual audit of motif-level explanations in cell-conditioned multi-omics drug response prediction: pipeline, processed GDSC2 x DepMap data and results

Evidence-integrating pipeline that nominates novel, druggable small-molecule cancer targets from DepMap dependency, synthetic lethality, single-cell specificity, safety, and tractability — with an LLM nomination ensemble and a gene-masking bias control. Validated on glioblastoma.

Machine learning pipeline for identifying ivermectin-associated biomarkers and drug repurposing opportunities across human cancers.

Local-first computational cancer-research platform: reproducible dependency studies, pre-registered guardrails, adversarially reviewed reports. Hypothesis generation, not clinical claims.

DepMap-based FAM50A dependency analysis in colorectal cancer with TCGA projection and CRISPR knockdown simulation.

Reproducible reconstruction of the broken-(TA)n repeat phenotype associated with WRN dependency in MSI cancer models using current public DepMap WGS.