Institute of Biotechnology, Life Sciences Center, Vilnius University, Vilnius, Lithuania.
Bioinformatics. 2018 Jun 15;34(12):2037-2045. doi: 10.1093/bioinformatics/bty048.
Protein sequence alignment forms the basis for comparative modeling, the most reliable approach to protein structure prediction, among many other applications. Alignment between sequence families, or profile-profile alignment, represents one of the most, if not the most, sensitive means for homology detection but still necessitates improvement. We aim at improving the quality of profile-profile alignments and the sensitivity induced by them by refining profile-profile substitution scores.
We have developed a new score that represents an additional component of profile-profile substitution scores. A comprehensive evaluation shows that the new add-on score statistically significantly improves both the sensitivity and the alignment quality of the COMER method. We discuss why the score leads to the improvement and its almost optimal computational complexity that makes it easily implementable in any profile-profile alignment method.
An implementation of the add-on score in the open-source COMER software and data are available at https://sourceforge.net/projects/comer. The COMER software is also available on Github at https://github.com/minmarg/comer and as a Docker image (minmar/comer).
Supplementary data are available at Bioinformatics online.
蛋白质序列比对是比较建模的基础，比较建模是蛋白质结构预测最可靠的方法之一，在许多其他应用中也是如此。序列家族之间的比对，或者说是轮廓-轮廓比对，是同源性检测中最敏感的方法之一，如果不是最敏感的方法，也需要改进。我们旨在通过细化轮廓-轮廓替换分数来提高轮廓-轮廓比对的质量和它们所引起的敏感性。
我们开发了一种新的评分方法，它代表了轮廓-轮廓替换评分的一个附加组成部分。全面的评估表明，新的附加评分在 COMER 方法的敏感性和对齐质量方面都有统计学上的显著提高。我们讨论了为什么这个评分会导致改进，以及它几乎最优的计算复杂度，这使得它可以很容易地在任何轮廓-轮廓对齐方法中实现。
该附加评分的实现已经在开源的 COMER 软件中可用，并可在 https://sourceforge.net/projects/comer 上获得。COMER 软件也可在 Github 上的 https://github.com/minmarg/comer 以及 Docker 镜像（minmar/comer）上获得。
补充数据可在 Bioinformatics 在线获得。
