Computational prediction of coiled-coil interaction structure specificity

Abstract

The alpha-helical coiled coil is a protein sequence and structural motif that consists of two or more helices in a parallel or antiparallel orientation supercoiling around a central axis. Coiled coils have been observed in a wide range of protein families, and many studies have focused on their sequence and structural diversity over the past half-century. In particular, the observation that coiled coils can be involved in determining protein-protein interactions and protein architectures has prompted the developments of methods to predict the structure of a coiled-coil complex from sequence information alone. In this thesis, I discuss the development of a structurally annotated database of coiled-coil sequence useful for training statistics-based methods of coiled-coil structure prediction. This database was used to retrain and stringently cross-validate the Multicoil method of predicting coiled-coil oligomerization state. In addition, I describe recent work using implicit and explicit structure models to predict dimeric coiled-coil orientation and alignment. Improvements to existing models, insight into coiled-coil structure determinants, and the future of coiled-coil prediction are also discussed.