Preclinical data lifts drug-response prediction in 27 of 38 compounds
david_van_dijk · x · 2026-10-28
CT-Translate was trained with and without preclinical data to predict patient cell drug response. Adding preclinical data improved patient AUC for 27 of 38 compounds (6 individually significant); 11 declined, none significantly. Outcome: ex vivo sensitivity of patient leukemia cells, retrospective five-fold cross-validation.
More from Apps
- AI designs 35,000 antibodies in weeks, opening new paths for food allergy treatment — SnackSafely · 2026-10-28
- Simon Willison: October 2026 is the vibe-coding-games inflection point, Opus 5.5 knows fun — Simon Willison · 2026-10-28
- Google tests query bubbles atop AI Overviews that funnel users into AI Mode — gaganghotra_ · 2026-10-28
- Community curates Awesome GPT-6.1 Terra list: projects, prompts and independent evals — Awacer · 2026-10-28
- Google's XR Language Teacher runs fully offline on your phone—is it a Duolingo killer? — kondasviktor · 2026-10-28
- 98-year-old who started driving in the Model A era buys first Tesla with FSD and calls it magic — XFreeze · 2026-10-28