The abstract is read at the publisher; the summary is JClub's.
A transcription factor activity representation learned by PRECISION from PRISM screen data predicts drug sensitivity in breast cancer cohorts, linking 623/1,447 drugs in SCAN-B to overall survival.
- Why it matters: Triple-negative breast cancer has limited targeted options, so drug repurposing needs a shared, interpretable way to generate hypotheses across cell lines and patient cohorts.
- What they did: PRECISION trains a heterogeneous GNN over prior-knowledge edges to model cell lines by TF activities, then uses Integrated Gradients to identify ten core TFs and per-drug Ridge models for cross-cohort prediction.
- The result: Predicted sensitivity is associated with overall survival for 623/1,447 drugs in SCAN-B and 74 in METABRIC, partly via a shared prognostic axis, and an explicit rule highlights seven kinase/HDAC inhibitors with prognostic associations that persist only partly in TNBC.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.