Beyond the Primary Target: How DeepTarget Rewrites Mechanism Discovery in Cancer

The modern small‑molecule rarely acts through a single, invariant target, and DeepTarget is a new computational framework that embraces this reality by mapping both direct and indirect anti‑cancer mechanisms across diverse cellular contexts. Developed by researchers at Sanford Burnham Prebys and collaborators, DeepTarget integrates large‑scale drug viability screens, genome‑wide CRISPR knockout data, and matched omics profiles to identify primary targets, context‑specific secondary targets, and mutation‑specific preferences that govern real‑world drug responses.​

DeepTarget uses a simple idea: if knocking out a gene with CRISPR makes cells respond like they do to a drug, that gene is likely the drug’s target or part of the pathway the drug disrupts.​

Because it looks at how living cells actually behave, not just how a drug might bind, it can uncover pathway and context effects that old “on‑target/off‑target” labels miss, opening more chances to repurpose drugs.

The study assembled a comprehensive foundation from DepMap, spanning 1,450 drugs and 371 cancer cell lines, and quantified drug–gene similarity with a Drug‑KO Similarity (DKS) score that explicitly adjusts for CRISPR and screen confounders. Using eight curated gold‑standard datasets of high‑confidence drug–target pairs, DeepTarget outperformed recent structure‑oriented tools in 7 of 8 benchmarks.

Source.