Gene coancestry in pedigrees and populations

Related individuals share common ancestors, and hence may carry DNA that is identical by descent (ibd) from these ancestors. With high probability, ibd DNA is of the same allelic type, leading to trait similarities among relatives. Classically, data on known relatives are used to map the genes underlying genetically mediated traits, and the prior probabilities of ibd are then given by the pedigree structure. However, pedigree data are expensive and difficult to collect, and the limited number of meioses within a set of known pedigrees leads to a lack of resolution in gene mapping. When pedigrees are ascertained for extreme trait values or from small populations, there are likely to be unknown relationships among the founder members of the same or of different pedigrees. Modern dense informative genetic marker data permit inference of ibd resulting from these unknown relationships, and this inferred ibd may be combined with ibd imputed within pedigrees to increase both the power and the resolution of mapping of genes contributing to complex quantitative traits. In this paper, we consider first the analysis of data within pedigrees, in terms of the ibd graph. This graph, defined among observed individuals and across the genome, specifies the segments of genome shared ibd among these individuals. Once the ibd graph is known, analyses of trait data may be carried out conditionally on the graph, and the pedigree relationships and genetic marker data are no longer relevant. We then show how ibd resulting from unknown more remote relationships can be estimated using a population-genetic based ibd model. Merging of the ibd graphs inferred within and among pedigrees provides a combined ibd graph, which may be used for trait-data analyses. We illustrate these methods with a small simulated-data example. We first examine the effect of genetic marker density on the inference of ibd in an extended pedigree. We then remove knowledge of some ancestors to create small subpedigrees, and analyze the ibd within and between these subpedigrees. Using the subpedigrees alone, linkage information is lost, but it is almost fully regained by inference of ibd among the subpedigrees. Software implementing these methods is available in the MORGAN-3 package (MORGAN V3.0.1 2010).