Ensemble Docking from Homology Models

• Published in: Journal of chemical theory and computation Vol. 6; no. 8; pp. 2547 - 2557
• Main Authors: Novoa, Eva Maria, Pouplana, Lluis Ribas de, Barril, Xavier, Orozco, Modesto
• Format: Journal Article
• Language: English
• Published: United States American Chemical Society 10.08.2010
• Subjects: Structure Prediction
• ISSN: 1549-9618; 1549-9626
• DOI: 10.1021/ct100246y

We present here a systematic exploration of the quality of protein structures derived from homology modeling when used as templates for high-throughput docking. It is found that structures derived from homology modeling are often similar in quality for docking purposes than real crystal structures, even in cases where the template used to create the structural model shows only a moderate sequence identity with the protein of interest. We designed an “ensemble docking” approach based on the use of multiple homology models. The method provides results which are usually of better quality than those expected from single experimental X-ray structures. The use of this approach allows us to increase around five times the universe of use of high-throughput docking approaches for human proteins, by covering over 75% of known human therapeutic targets.