| Literature DB >> 27159634 |
Sumudu P Leelananda1, Robert L Jernigan2,3, Andrzej Kloczkowski1,4.
Abstract
Highly designable structures can be distinguished based on certain geometric graphical features of the interactions, confirming the fact that the topology of a protein structure and its residue-residue interaction network are important determinants of its designability. The most designable structures and least designable structures obtained for sets of proteins having the same number of residues are compared. It is shown that the most designable structures predicted by the graph features of the contact diagrams are more densely packed, whereas the poorly designable structures are more open structures or structures that are loosely packed. Interestingly enough, it can also be seen that the highly designable identified are also common structural motifs found in nature.Keywords: contact maps; designability; graph features; interaction network; lattice models; machine learning; network; prediction; structure
Mesh:
Substances:
Year: 2016 PMID: 27159634 PMCID: PMC4876523 DOI: 10.1089/cmb.2015.0209
Source DB: PubMed Journal: J Comput Biol ISSN: 1066-5277 Impact factor: 1.479