Modof is a generative AI tool for lead optimization. Given a lead to optimize and its desired properties, Modof intelligently changes a few fragments of the lead molecule and modifies the lead into another similar molecule which has better desired properties. Modof generates such optimized molecules automatically without expert knowledge input. 

Referece: PMID:35187404

Contact: Xia Ning

G2retro is a generative AI tool to generate one-step synthesis routes for any given product. It can produce multiple possible synthesis routes that are feasible and can vary. G2retro generates such routes without expert knowledge input. 

Reference: PMID: 37253928

Contact: Xia Ning

TCRPPO is a generative AI tool for peptide design. It generates T-cell receptors (TCRs) that can better recognize given peptides. 

Reference: https://link.springer.com/chapter/10.1007/978-3-031-29119-7_11

Contact: Xia Ning

LlaSMol is a Large-Lanauge Model (LLM) together with a high-quality instruction tuning dataset for a wide variety of tasks in drug development, including property prediction, forward synthesis, name convension, etc. 

Reference: to appear in COLM 2024

Contact: Xia Ning

Deep learning framework for high-throughput mechanism-driven phenotype compound screening

Reference: PMID: 33796820

Contact: Ping Zhang

Chemical-induced gene expression ranking.

Reference: PMID: 35465231

Contact: Ping Zhang