Daily Digest

#artificialintelligence 

Chemical probes are important tools for understanding biological systems. However, because of the huge combinatorial space of targets and potential compounds, traditional chemical screens cannot be applied systematically to find probes for all possible druggable targets. Here, researchers demonstrate a novel concept for overcoming this challenge by leveraging high-throughput metabolomics and overexpression to predict drug–target interactions. The metabolome profiles of yeast treated with 1,280 compounds from a chemical library were collected and compared with those of inducible yeast membrane protein overexpression strains. By matching metabolome profiles, they predicted which small molecules targeted which signaling systems and recovered known interactions.