TOOLS FOR COMPARATIVE GENOME ANALYSIS

Roslin Institute (Edinburgh) Roslin Midlothian EH25 9PS Scotland, U.K.

It is widely accepted that comparative genome mapping will provide livestock genome mappers a framework to exploit knowledge of the human genome in searches for livestock trait genes. Bioinformatics tools and resources are required to achieve this objective. We are developing such tools for this purpose. Criteria have been developed for evaluating whether genes characterised in different species are homologous or not. These criteria include the degree of nucleic acid (or encoded protein) sequence similarity, patterns of gene expression, functional similarity and relative map location. The primary objective for this project is to develop systems for assessing homology states on the basis of these criteria. Although there are multiple criteria there is only a significant body of data for two - sequence similarity and relative map location. Thus, our initial system has addressed these two criteria. For species A and B we consider genes for which there is an associated SWISS-PROT entry and a map location. The genes are ordered by using a mid-point model of map location as many of the map locations are defined as overlapping regions. For each gene in species we search for sequence matches using BLAST and consider sequence match, relative sequence match and relative sequence length similarity and a matrix based on these features. An algorithm has been developed to cluster the candidate homologous genes using the map information. Those genes that do cluster are considered to be orthologues. We have evaluated these procedures for comparisons of genes across species and more recently for comparing ESTs across species. Results from the homology assessment procedures have been used to initiate population of The Comparative Animal Genome database (TCAGdb) (see http://www.ri.bbsrc.ac.uk/bioinformatics/).