Assessing Predicted Contacts for Building Protein Three-Dimensional Models.

ABSTRACT: Recent successes of contact-guided protein structure prediction methods have revived interest in solving the long-standing problem of ab initio protein structure prediction. With homology modeling failing for many protein sequences that do not have templates, contact-guided structure prediction has shown promise, and consequently, contact prediction has gained a lot of interest recently. Although a few dozen contact prediction tools are already currently available as web servers and downloadables, not enough research has been done towards using existing measures like precision and recall to evaluate these contacts with the goal of building three-dimensional models. Moreover, when we do not have a native structure for a set of predicted contacts, the only analysis we can perform is a simple

SUBMITTER: Adhikari B 

PROVIDER: S-EPMC5512107 | biostudies-literature | 2017

REPOSITORIES: biostudies-literature