UNIT 3.4 Finding Homologs in Amino Acid Sequences Using Network BLAST Searches

Abstract

The Basic Local Alignment Search Tool (BLAST) is the most fundamental (and most misused) algorithm and software in bioinformatics/computational biology for functional assessment of unknown proteins or discovery of similar proteins with potentially common evolutionary origins. We show how to balance sensitivity with selectivity (without generating massive output) by selecting and demonstrating proper database, algorithm, and alignment display options of the user-friendly Web sites of the National Center for Biotechnology Information (NCBI). We discuss protein query searches against protein databases and submission of all combinations of translated searches. Careful biological and statistical inferences are drawn to possible functions, taking into account the highly nonrandom nature of proteins. Guidelines for such inferences, using real-life biological examples (e.g., protein kinases with widely distributed structural and functional domains), are provided. We show how to avoid incorrect functional inference from misleading similarities, using the divergent evolution of a serine protease domain that erodes the protease function in haptoglobins. Curr. Protoc. Bioinform. 25:3.4.1-3.4.34. © 2009 by John Wiley & Sons, Inc.