Utilizing twilight zone sequence similarities to increase the accuracy of protein 3D structure comparison
Alignment algorithms, Protein structure, Structural alignment
ACM BCB 2014 - 5th ACM Conference on Bioinformatics, Computational Biology, and Health Informatics
Protein structure alignment is often used to infer a sequential alignment. The inverse problem of utilizing alignments of protein sequences to derive structural alignments has been extensively studied, but the applications are still scarce. Here we present pMatch - A novel algorithm that takes advantage of sequential profiles to compute extremely accurate protein structure alignments. Contrary to some claims that sequential information is only useful when aligning structures of proteins that share a high degree of sequence identity, we demonstrate that incorporating sequential information can increase accuracy of alignments even between proteins sharing only about 10% of identical residues. We use two widely established benchmarks to show that our method is capable of accurately aligning protein structures that exhibit substantial conformational changes, large residue insertions/deletions and circular permutations. A preliminary version of the pMatch Web server is available at http://bioinformatics.cs.uni.edu.
Original Publication Date
DOI of published version
Poleksic, Aleksandar and Gray, Paul, "Utilizing twilight zone sequence similarities to increase the accuracy of protein 3D structure comparison" (2014). Faculty Publications. 1361.