Automated evaluation of quaternary structures from protein crystals

• DOI: 10.1371/journal.pcbi.1006104
• Publication Date: 2018-04-30T13:34:49Z
• Authors: Spencer Bliven; Aleix Lafita; Althea Parker; Guido Capitani; Jose M. Duarte
• Keywords: Protein crystallization; Protein Data Bank; Interface (matter); Lattice (music)

A correct assessment of the quaternary structure proteins is a fundamental prerequisite to understanding their function, physico-chemical properties and mode interaction with other proteins. Currently about 90% structures in Protein Data Bank are crystal structures, which embedded lattice among number contacts. Computational methods required 1) classify all protein-protein contacts lattices as biologically relevant or 2) provide an how interfaces combine into biological assembly. In our previous work we addressed first problem EPPIC (Evolutionary Interface Classifier) method. Here, present solution second new method that combines interface classification results symmetry topology considerations. The algorithm enumerates possible valid assemblies within using graph representation predicts most probable unit based on pairwise scoring. Our achieves 85% precision (ranging from 76% for different oligomeric types) dataset 1,481 consensus PDB annotations. Although almost same achieved by PISA, currently popular assignment method, show that, due fundamentally approach problem, two complementary could be combined improve assembly assignments. software automatic protein (EPPIC version 3) has been made available through web server at http://www.eppic-web.org.