Iterative Hybrid Alignment: Improving the Sensitivity of Biological Database Searches

俄亥俄州立大学（Ohio State University）获得了一笔拨款，用于建立一个计算机程序，使研究人员能够在基因序列中找到匹配或匹配的区域，即使这些匹配非常弱。不同生物的蛋白质序列之间的匹配程度和匹配的细节是理解所有生物共同的进化和基本生物过程的核心。随着序列数据的数据库变得越来越大，问题变得越来越深远，对更智能的搜索和匹配算法的需求变得至关重要。在这个项目中，PI将开发一种高度敏感的搜索算法，即使在实际同源性非常弱的情况下也能识别同源性。所提出的算法比目前的算法更好地利用了插入和删除的计算统计成本，因为它们在序列中是不同的。第一个实现是作为PSI-BLAST的重新设计而提出的，但它相当于完全取代了流行的蛋白质序列搜索算法的核心算法。外围功能也将发生重大变化。这是一种新工具，将对涉及匹配算法的研究以及基因组学/蛋白质组学社区产生影响。

A grant has been awarded to Ohio State University to build a computer program that will allow researchers to find matches or matching areas in genetic sequences even when those matches are very weak. The degree of matching and the specifics of that matching among protein sequences of different organisms is central to understanding evolution and fundamental biological processes common to all organisms. As the databases of sequence data become larger and questions become more far reaching, the need for smarter search and matching algorithms becomes crucial. In this project, the PI's will develop a highly sensitive search algorithm for identifying homologies even in cases in which the actual homologies are very weak. The algorithm proposed makes much better use than current algorithms of the calculated statistical costs of insertions and deletions as they vary across the sequence. The first implementation is proposed as a redesign of PSI-BLAST but it amounts to a complete replacement of the core algorithm of that popular protein sequence search algorithm. There will also be significant changes in the peripheral functions as well. This is a new tool that will have impacts on research involving matching algorithms in general as well as within the genomics/proteomics community.