Optimization of a new score function for the detection of remote homologs

The growth in protein sequence data has placed a premium on ways to infer s tructure and function of the newly sequenced proteins. One of the most effe ctive ways is to identify a homologous relationship with a protein about wh ich more is known. While close evolutionary relationships can be confidentl y determined with standard methods, the difficulty increases as the relatio nships become more distant. All of these methods rely on some score functio n to measure sequence similarity. The choice of score function is especiall y critical for these distant relationships. We describe a new method of det ermining a score function, optimizing the ability to discriminate between h omologs and non-homologs. We find that this new score function performs bet ter than standard score functions for the identification of distant homolog ies. Proteins 2000;41:498-503. (C) 2000 Wiley-Liss, Inc.