Novel DNMT inhibitors discovered to date have all been knowledge-based and sampled an extremely small portion of chemical property space and compounds identified this way have had negligible clinical impact to date, suggesting that fresh approaches are needed. This project aims to utilize an unbiased readout to interrogate the full DNA methylation pathway against a broad panel of compounds in a high-throughput screen. During this period, the project team conducted quantitative high-throughput screening (qHTS) on select small molecule libraries, and hit compounds were validated. Machine learning approaches were utilized to construct in silico quantitative structure-activity relationship (QSAR) models to enable virtual screening of compounds against larger collections of NCATS molecules. The validation of hits in orthogonal assays is currently underway.
