Arc Institute's PIE model taps decades of curated biology to predict unseen gene perturbation responses
anshulkundaje · x · 2026-10-06
Arc Institute released PIE, a perturbation-prediction model that goes beyond learning only from experiments. By incorporating decades of curated biological knowledge, PIE can predict which genes respond to a perturbation — even in biological contexts and for perturbations it has never seen before.
More from Research
- Virtual cell debate was a metrics problem: Nature Biotech paper says calibrated metrics vindicate perturbation models — anshulkundaje · 2026-10-06
- First Oracle PL/SQL benchmark, ported from MBPP/MBPP+, to be presented at EMNLP — srchvrs · 2026-10-06
- EvalEval Coalition on the science of AI evaluation grows to 800+ researchers, adopts new governance — evijit · 2026-10-06
- Calico's Multi-ISM cuts genomic saturation mutagenesis evaluations 45-fold — anshulkundaje · 2026-10-06
- SkillGym: Fine-Tuning on Verified Skill Runs Lifts Terminal-Bench 2.1 Success by 19 Points — rohanpaul_ai · 2026-10-06
- 14,000+ PhD Placement Records From 80 Econ Departments Now Searchable in Open Dataset — Afinetheorem · 2026-10-06