LabCompass preprint uses generative inverse design to steer human hematopoietic cell fate in vitro

anshulkundaje · x · 2026-10-09

• Fabian Theis and colleagues released a new bioRxiv preprint, LabCompass, applying generative inverse design to a key question: which experiment yields the cells we want?
• The team previously worked on predicting how cells respond to perturbations; this work flips the problem toward actively designing perturbation experiments, where combination and dose spaces are too large to exhaustively screen.
• Method: LabCompass couples a generative model of experimental designs with a forward model of cellular responses, using a desired phenotype to guide generation across a continuous design space — including in unexplored regions where predictive models are weakest.
• Validated by steering human hematopoietic cell fate in vitro; the authors call inverse design one of the most exciting directions for the field.
• Preprint (not yet peer reviewed): doi.org/10.64898/2026.10.07.756550

Original post →

More from Research

• AI2's Nature Paper: Byteification Retrofits LLMs to Byte-Level for Under 1% of Pretraining Cost — TheTuringPost · 2026-10-09
• Ramanujan parallel: mathematicians call AI proofs a threat, not genius — CatAstro_Piyush · 2026-10-09
• Datology AI: Curated Pretraining Data Lifts 30B-A3B to 46.8% vs 37.7% Baseline — josh_wills · 2026-10-09
• Ai2 rebuilt its GPU scheduler: median queue wait fell from 5 minutes to 24 seconds on H100 cluster — allen_ai · 2026-10-09
• New COLM Workshop Paper Measures and Reduces Slop in Long-Horizon Coding Agents — dan_fried · 2026-10-09
• Cryptographer Lays Out a Balanced Take on AI's Impact on Cryptography — jedisct1 · 2026-10-09

Research channel →