A method for increasing virtual screening success for non-experts

What is it about?

Although virtual screening methods play a vital role in contemporary drug discovery, their performance usually demonstrates striking fluctuations that hinder error-free use and reveal extensive pre-optimization requirements. The present analysis introduces a conceptual framework for rationally combining screening tools characterized by orthogonality into a consensus scheme enabling optimal exploitation of individual method capacities. Results show that the consensus approach succeeds in introducing a considerable degree of stability on screening performance at reasonable computational cost, whereas individual screens seem to be heavily dependent on starting conditions. The benchmark is coupled with an analysis of residuals converging to the notion that consensus ranking comprises a valid tool with markedly improved robustness and thus highly suitable for prospective screening. Finally, screening reveals a potent inhibitor of DYRK1α kinase and a cell-based phosphoproteomic analysis of its effects in key signaling pathways and regulatory modules is reported.

Featured Image