Defining the Role of Essential Genes in Human Disease

Christos A Ouzounis,Ana Zhu,Kathryn E Hentges,David L Robertson,Jonathan E Dickerson

doi:10.1371/journal.pone.0027368

Abstract

A greater understanding of the causes of human disease can come from identifying characteristics that are specific to disease genes. However, a full understanding of the contribution of essential genes to human disease is lacking, due to the premise that these genes tend to cause developmental abnormalities rather than adult disease. We tested the hypothesis that human orthologs of mouse essential genes are associated with a variety of human diseases, rather than only those related to miscarriage and birth defects. We segregated human disease genes according to whether the knockout phenotype of their mouse ortholog was lethal or viable, defining those with orthologs producing lethal knockouts as essential disease genes. We show that the human orthologs of mouse essential genes are associated with a wide spectrum of diseases affecting diverse physiological systems. Notably, human disease genes with essential mouse orthologs are over-represented among disease genes associated with cancer, suggesting links between adult cellular abnormalities and developmental functions. The proteins encoded by essential genes are highly connected in protein-protein interaction networks, which we find correlates with an over-representation of nuclear proteins amongst essential disease genes. Disease genes associated with essential orthologs also are more likely than those with non-essential orthologs to contribute to disease through an autosomal dominant inheritance pattern, suggesting that these diseases may actually result from semi-dominant mutant alleles. Overall, we have described attributes found in disease genes according to the essentiality status of their mouse orthologs. These findings demonstrate that disease genes do occupy highly connected positions in protein-protein interaction networks, and that due to the complexity of disease-associated alleles, essential genes cannot be ignored as candidates for causing diverse human diseases.

Highlights

Much effort has been invested in identifying the set of genes that when mutated have a causal relationship with human disease
To determine if differences exist between essential and nonessential disease genes we examined several parameters in our analyses, including the physiological systems affected by each disease gene, the connectivity of each gene in protein-protein interaction networks, and the genetic mechanisms by which genes cause human disease
To assess whether essentiality was correlated with particular disease gene properties, we grouped the disease genes into viable and lethal categories, based on inference from mouse knockout data [26,27]

Summary

Introduction

Much effort has been invested in identifying the set of genes that when mutated have a causal relationship with human disease. Further studies that classified disease genes based on their requirement during development, or essentiality, led to the conclusion that the majority of disease genes are non-essential [5,6,7,8]. This conclusion is drawn from the analysis of human disease genes based on the phenotypes of their mouse orthologs. Disease genes whose mouse orthologs produce lethal phenotypes when deleted were considered essential, and all other genes considered nonessential This classification, over-estimates the size of the non-essential gene group, due to the inclusion of genes with no reported knockout data. To include genes with no knockout data in the viable gene group confounds the analysis, and could lead to erroneous conclusions about the relative importance of lethal and viable genes in human disease

Methods

Results

Conclusion