Homozygosity, inbreeding and health in European populations

Abstract

Inbreeding results in increased levels of homozygosity for deleterious recessive alleles, leading to increased incidence of monogenic disease in inbred families. It has also been suggested that inbreeding increases the risk of diseases such as cancer and heart disease, implying a role for the combined effects of many recessive alleles distributed across the genome. A better understanding of the links between inbreeding, homozygosity and disease is therefore of interest to those concerned with understanding the genetic architecture of complex disease. A homozygous genotype is defined as autozygous if both alleles originate from the same ancestor. Quantifying inbreeding involves quantifying autozygosity. A new, observational method of quantifying autozygosity using genomic data is developed here. Based on runs of homozygosity (ROH), this approach has a sound theoretical basis in the biological processes involved in inbreeding. It is also backed by strong empirical evidence, correlating strongly with pedigree-derived estimates of inbreeding and discriminating well between populations with different demographic histories. ROH are a signature of autozygosity, but not necessarily autozygosity of recent origin. Short ROH are shown to be abundant in demonstrably outbred individuals and it is suggested that this is a source of individual genetic variation which merits investigation as a disease risk factor, although denser genotype scans than those used in the present study are required for the reliable detection of very short ROH. In the absence of such dense scans, it is suggested that ROH longer than 1 or 1.5 Mb be used to estimate the effects of inbreeding on disease or quantitative physiological traits (QT), and that a simple measure of homozygosity be used to investigate overall recessive effects. Evidence for recessive effects on 13 QT important in cardiovascular and metabolic disease was investigated in 5 European isolate populations, characterised by heightened levels of inbreeding. A significant decrease in height was associated both with increased homozygosity and (to a lesser extent) with increased ROH longer than 5 Mb (i.e. inbreeding) estimated using a 300,000 SNP panel. No evidence was found for recessive effects on any of the other QTs. Evidence for recessive effects on colorectal cancer risk were investigated in two outbred case control samples typed with a 500,000 SNP panel. Cases were significantly more homozygous and had more of their genome in short ROH than did controls. Cases were significantly more homozygous than controls even when inbred individuals were removed from the sample. There was also some evidence of an inbreeding effect, with inbred subjects having slightly significantly higher odds of colorectal cancer than outbred subjects. This study provides evidence of recessive effects on a common, complex disease in outbred populations and on height in both inbred and outbred populations and shows that such effects are not solely attributable to increased levels of homozygosity resulting from recent inbreeding. Individual variation among outbred individuals in the proportion of the genome that is homozygous may be important in disease risk. The development of denser genotype scans will facilitate better enumeration of short ROH in outbred individuals so that these can be properly enumerated and investigated as a disease risk factor.