Output details
11 - Computer Science and Informatics
University College London
Accurate de novo structure prediction of large transmembrane protein domains using fragment-assembly and correlated mutation analysis.
<28>This paper describes a breakthrough in predicting structures for transmembrane proteins, which are of huge biomedical importance; TM proteins being targets of most drugs. As experimental techniques usually fail for TM proteins, biologists rely heavily on computational methods. For the very first time, we have demonstrated how the 3-D structures of large TM proteins can be modelled accurately from sequence information alone. The key novelty here is a new stochastic recombination algorithm for generating accurate 3-D protein models from sparse noisy distance constraints. This algorithm will help reveal the structures of hundreds of hitherto intractable proteins of major biomedical significance.