The Genome Comparison Project: Improving protein functional annotation in databases

Over the years, a rather large body of secondary information (structural, functional, similarities to other entries and a variety of cross-references) has been attached to protein database entries. Once such information is entered, it rarely gets updated or corrected. Thus, annotation of predicted protein function is often incomplete, uses non-standardized nomenclature or can be incorrect when inferred from previous, incorrectly annotated sequences. Additionally, many proteins are composed of several structural and/or functional domains (modules comprising distinct evolutionary, functional and structural units), which can be overlooked by automated annotation procedures. Moreover, the comparative information available today is huge when compared to the early days of genomics.

The main objective of the Genome Comparison Project is to perform a complete pairwise comparison between all predicted protein sequences, obtaining similarity indices that will be used, together with standardized Gene Ontology (http://www.geneontology.org), as a reference repository for the annotator community, providing an invaluable data source for biologists. The sequence similarity comparison program used in the Genome Comparison Project is called SSEARCH (W.R. Pearson [1991] Genomics 11:635-650), a freely available implementation of the Smith-Waterman rigorous algorithm (T. F. Smith and M. S. Waterman [1981] J. Mol. Biol. 147:195-197) (algorithm is an organized procedure for performing a given type of calculation or solving a given type of problem), which finds the mathematically best local alignment between pairs of sequences.

As a result, precise annotation, correction of inconsistencies, and assignment of possible functions to hypothetical proteins of unknown function will be possible. Moreover, proteins with multiple domains and functional elements will be correctly spotted. Even distant relationships will be detected.