Experimental screening of dihydrofolate reductase yields a "test set" of 50,000 small molecules for a computational...

High-throughput screening (HTS) generates an abundance of data that are a valuable resource to be mined. Dockers and data miners can use "real-world" HTS data to test and further develop their tools. A screen of 50,000 diverse small molecules was carried out against Escherichia coli dihydrofolate reductase (DHFR) and compared with a previous screen of 50,000 compounds against the same target. Identical assays and conditions were maintained for both studies. Prior to the completion of the second screen, the original screening data were publicly released for use as a "training set", and computational chemists and data analysts were challenged to predict the activity of compounds in this second "test set". Upon completion, the primary screen of the test set generated no potent inhibitors of DHFR activity.