Recently, Professor Lan Yanyan from the Institute for Intelligent Industry (AIR) of Tsinghua University collaborated with teams from the School of Life Sciences and the Department of Chemistry to successfully develop a high-throughput drug virtual screening platform called DrugCLIP, which is driven by artificial intelligence. This innovative achievement was published in the journal Science on January 9, 2026, with the title "Deep Contrastive Learning Enables Drug Virtual Screening at the Genomic Level."

In the current field of drug development, only about 10% of human druggable targets have been explored. Facing thousands of potential targets, how to quickly identify effective compounds in the vast chemical space has become a major challenge. The introduction of DrugCLIP effectively solves this problem, improving the screening speed by a million times compared to traditional methods, while also achieving significant improvements in prediction accuracy.

Based on the DrugCLIP platform, the research team completed the first drug virtual screening covering the scale of the human genome. This process involved approximately 10,000 protein targets and 20,000 protein pockets, and analyzed and screened more than 500 million drug-like small molecules, ultimately enriching over 2 million potential active molecules. In addition, the protein-ligand screening database built by the team has become the largest database of its kind and has been freely opened to the global scientific community, greatly facilitating the work of researchers.

Key Points:

🌍 The DrugCLIP platform improves drug screening speed by a million times, helping the development of targeted drugs.  

🔬 The study covers the scale of the human genome, screening over 2 million potential active molecules.  

📚 The protein-ligand screening database is now freely available to the global scientific community.