Role of duplicate genes in robustness against deleterious human mutations

Abstract

It is now widely recognized that robustness is an inherent property of biological systems [1],[2],[3]. The contribution of close sequence homologs to genetic robustness against null mutations has been previously demonstrated in simple organisms [4],[5]. In this paper we investigate in detail the contribution of gene duplicates to back-up against deleterious human mutations. Our analysis demonstrates that the functional compensation by close homologs may play an important role in human genetic disease. Genes with a 90% sequence identity homolog are about 3 times less likely to harbor known disease mutations compared to genes with remote homologs. Moreover, close duplicates affect the phenotypic consequences of deleterious mutations by making a decrease in life expectancy significantly less likely. We also demonstrate that similarity of expression profiles across tissues significantly increases the likelihood of functional compensation by homologs.

Figure 1. The relationship between the sequence identity of the closest homolog and the conditional probability of a disease gene, P(disease|sequence_identity_of_closest_homolog).

Genes with close paralogs are less likely to harbor disease mutations. For display purposes, we assumed that 20% of all human genes harbor disease mutations (see Methods). The sets of all human genes used for the calculations are A) Ensembl and B) Swiss-Prot.

Figure 2. The relationship between the sequence identity of the closest homolog and the ratio of non-synonymous to synonymous human SNPs per site (Ka/Ks).

The Ka/Ks ratio was averaged for genes within each sequence identity bin. The ratio is shown for all (black) and only for validated (red) SNPs from the dbSNP database . Above 60% sequence identity, the Ka/Ks ratio increases monotonically as the homolog sequence identity increases.

Figure 3. Influence of the close duplicates on disease phenotypes.

The phenotypic disease data (reduction in life expectancy) were obtained from the study by Jimenez-Sanchez et al. . For display purposes, we show the proportion of genes with close duplicates (sequence identity to the closest paralog > = 60%) in each phenotype bin. The proportion of genes with close duplicates decreases with the reduction in life expectancy.

Figure 4. The Similarity of Tissue Expression (STE) increases the likelihood of functional compensation.

The STE value (Jaccard's coefficient of similarity for gene expression patterns) reflects the similarity of expression between duplicates across tissues. The average STE was calculated for gene pairs within each sequence identity bin. The average STE is consistently lower for disease genes (black) compared to all genes (red) (see also Table 5S, Supporting Information).