Pengtao Xie

Our group develops multi-modal large language models for predicting protein functions and properties, as well as the indications, pharmacodynamics, and mechanisms of action of candidate drug compounds. We develop multi-modal foundation models on single-cell transcriptomics data for gene perturbation prediction, cancer drug response prediction, gene regulatory network inference, etc. Additionally, we develop multi-modal foundation models on protein sequences, structures, and literature texts for protein function and property prediction.