🔬Causal Models Need Causal Data - Xaira’s X-Cell model for Drug Discovery (Bo Wang & Ci Chu, Chief Discovery Officer & Chief AI Scientist)
AI drug discoveryCausal modelsGene expression predictionHigh-throughput experimentationVirtual cell modelsProtein designPatient stratificationSingle-cell genomicsCRISPR perturbationDiffusion modelsBiological prior knowledgeModel generalizationOpen scienceSpatial transcriptomicsTemporal cell dynamics
This episode features Bo Wang and Si Chu from Xaira Therapeutics, discussing their AI drug discovery platform, X-Cell, which uses high-throughput experimentation to generate causal data for predicting cellular responses to genetic perturbations. They explain how X-Cell, a diffusion language model, outperforms linear baselines in generalizing predictions to unseen cell types and contexts, addressing critical data limitations in drug discovery. The conversation also delves into the importance of open science, the evolving role of academics versus industry in AI research, and future technological breakthroughs needed for comprehensive virtual cell models.