Large-scale search for genes on which positive selection may operate

There has been great controversy between neutralists and selectionists concerning what proportion of mutants contributed to gene evolution as advantageous mutants. Several studies have shown that there are some genes on which positive selection operates through comparison and statistical analysis of nucleotide sequences, but no one have been able to estimate the proportion of genes on which positive selection operates among all the genes available. One reason is that conducting such research involves a large amount of DNA sequence data and coordinated analysis in order to conduct large-scale searches for candidate genes making full use of sophisticated computers. The purpose of this study was to investigate the role of positive selection in the course of gene evolution. For the first attempt, I estimated the proportion of genes on which positive selection may operate utilizing a large-scale search. In PART I, I discuss the strategy and results of a large-scale search for genes on which positive selection may operate. In this study, a large-scale , search is defined as search for genes which can fulfill certain criteria by utilizing a large amount of DNA sequence data using the computers of high level processibility including a super computer. The search was performed using the following strategies: 1) alignment construction for the 3,595 groups of homologous sequences of non-partial coding sequences available in the DDBJ DNA Database Release 11; 2) comparison of numbers of synonymous and non synonymous substitutions. From the search, 17 homologous gene groups were obtained as candidate genes on which positive selection may operate. They constituted 0.47% of all the gene groups examined in the study. Interestingly enough, nine out of the 1 7 gene groups were for surface proteins of viruses and parasites. A window analysis of numbers of synonymous and nonsynonymous substitutions was performed to locate possible within-gene regions subject to positive selection. The analysis indicated that a region where the number of nonsynonymous substitutions was larger than that of synonymous substitutions corresponded to a known antigenic epitope region. In PART II, an additional search was conducted to find genes which contained a region or a segment subject to positive selection. This search was performed because the nucleotide sites responsible for positive selection usually are not spread throughout the entire gene even when positive selection is operating. This was the cause of my inability to locate MHC genes in the search in PART I, because positive selection is shown to be operating only on the antigenic sites in MHC genes. In the new search shown in part II, 192 gene groups were obtained as candidate genes having within-gene regions where positive selection may operate. They constituted about five percent of all the genes examined.