In silico screening of drug targets against thousands of select compounds
Leslie Klis McNeil, Matthew Cohoon, Kaitlyn Hwang, Michael Kubal, Gohar Rem Margaryan, Bruce Parrello, Fangfang Xia, Jenifer F. Zinner, Rick L. Stevens · The FASEB Journal · 2007
The National Microbial Pathogen Data Resource (NMPDR, www.nmpdr.org ) identifies candidate drug targets and virtually screens potential drugs. NMPDR is one of eight Bioinformatic Resource Centers (BRC) established by NIAID to support research on potential agents of biowarfare, bioterrorism, or emerging disease. Candidate targets are essential for growth or virulence in at least one NMPDR focus organism, have been included in biological subsystems by NMPDR curators, have orthologs with experimentally determined structures in Protein Data Bank, and have orthologs in a substantial number of bacterial Priority Pathogens curated by the BRCs. In silico screening is the computational molecular docking of a library of ligands against a protein structure using a force field method to compute the binding energy. Targets are first docked with 1.5K randomly selected compounds. The results are used as a training set for a neural network that evaluates the characteristics of binding and non‐binding compounds. Libraries of commercially available or FDA approved drug compounds are screened with the neural net to select an enriched set of probable ligands. Molecular docking is then performed with the selected compounds using Dock5 on a BlueGene supercomputer. The top 1000 docked compounds are posted on NMPDR as targets are screened. Funded by NIAID under Contract No. HHSN266200400042C.