Zero-shot biological reasoning with open-weights large language models reproduces CRISPR screen based prediction of synthetic lethal interactions.
Aurél György Prósz, Zsófia M. Sztupinszki, Miklós Dióssy, Oz Kilim, Bogumil Zimon, István Csabai, Zoltán Szállási
This paper uses open-weight large language models to predict synthetic lethal interactions, outperforming random chance and non-LLM methods in reproducing CRISPR screen results.
bioRxiv · May 11, 2026 · PreprintBioinformatics