Functional and evolutionary implications of gene orthology

Abstract

Orthologues and paralogues are types of homologous genes that are related by speciation or duplication, respectively. Orthologous genes are generally assumed to retain equivalent functions in different organisms and to share other key properties. Several recent comparative genomic studies have focused on testing these expectations. Here we discuss the complexity of the evolution of gene-phenotype relationships and assess the validity of the key implications of orthology and paralogy relationships as general statistical trends and guiding principles.

Figure 1. Decay of the number of one-to-one orthologues with the increase of the intergenomic evolutionary distance

Here we use bidirectional best hits (BBHs) as a proxy for one-to-one orthologues. The dependency of the fraction of genes that can be assigned a BBH on the evolutionary distance between genomes is shown for two genomes: the bacterium Escherichia coli K12 (a); and the archaeon Haloarcula marismortui (b). In this analysis, these are termed the ‘master’ genomes. For all proteins encoded in each of the master genomes, a BLASTP search was carried out against the protein sequences from 573 representative bacterial and archaeal genomes, and for the most similar proteins (that is, the best hits), a reciprocal BLASTP search was carried out to identify BBHs. The BLASTP score for each BBH was normalized by the self-hit score in the master genome and converted into distance using the formula distance = –ln(score). The distance between the genomes that were compared was estimated as the median distance between BBH pairs. Please see REF. for further details. The figure is modified, with permission, from REF. © (2012) Oxford Univ. Press.

Figure 2. Functional divergence versus sequence divergence for orthologues and paralogues

Plotted is the excess of functional similarity between orthologous pairs of genes as compared to paralogues for different degrees of sequence divergence and for different types of functional ontologies. The data are from REF. , and further details of the analysis can be found in that paper. Functional similarities are averaged over homologous gene pairs (as predicted by the OMA method) from 13 model species (namely, Homo sapiens, Mus musculus, Rattus norvergicus, Drosophila melanogaster, Danio rerio, Caenorhabditis elegans, Saccharomyces cerevisiae, Schizosaccharomyces pombe, Candida albicans, Dictyostelium discoideum, Arabidopsis thaliana, Escherichia coli and Pseudomonas aeruginosa). They are based on comparisons of Gene Ontology annotations backed by experimental evidence extracted from publications sharing no common authors. In addition, this measure takes into account annotation biases in the different species and other possible confounding factors. Nonsignificant differences between orthologues and paralogues (P value >0.001, using a Mann–Whitney U test) are indicated in grey.