Comparative modeling of proteins

Abstract

Much of the biochemistry that underlies health, medicine, and numerous biotechnology applications is regulated by proteins, whereby the ability of proteins to effect such processes is dictated by the threedimensional structural assembly of the proteins. Thus, a detailed understanding of biochemistry requires not only knowledge of the constituent sequence of proteins, but also a detailed understanding of how that sequence folds spatially. Three-dimensional analysis of protein structures is thus proving to be a critical mode of biological and medical discovery in the early twenty-fi rst century, providing fundamental insight into function that produces useful biochemistry and dysfunction that leads to disease. The large number of distinct proteins precludes rigorous laboratory characterization of the complete structural proteome, but fortunately effi cient in silico structure prediction is possible for many proteins that have not been experimentally characterized. One technique that continues to provide accurate and effi cient protein structure predictions, called comparative modeling, has become a critical tool in many biological disciplines. The discussion herein is an updated version of a previous 2008 treatise focusing on the general philosophy of comparative modeling methods and on specifi c strategies for successfully achieving reliable and accurate models. The chapter discusses basic aspects of template selection, sequence alignment, spatial alignment, loop and gap modeling, side chain modeling, structural refi nement and validation, and provides an important new discussion on automated computational tools for protein structure prediction.