Estimating relatedness with no prior specification of any genealogy: The role of the X-chromosome

Abstract

Inferential studies on genetic relatedness fall into two broad approaches: (a) the genetic types of the individuals are used to compare the likelihood of their relatedness through specific genealogies; or (b) the genetic information is evaluated to infer the degree (or level) of relatedness between the individuals without specifying any pedigree. Approach (a) is used to compute statistics when the alternative pedigrees and/or samples are few (such as in the case of a standard paternity test, for example), while approach (b) is more general, and useful to bring into light cryptic, undisclosed or unspecified kinships (suited, for example, to seek for related individuals at mass disasters comparing recovered and reference samples). From a theoretical point of view, framework (a) is based on the complete set of identical-by-descent probabilities (three parameters for autosomes and a pair of non-inbred individuals), while (b) aims, in most cases, at the estimation of a single parameter: the coancestry coefficient θ, which is generally assumed as a coarser measure of genetic relatedness since it clusters pedigrees, such as “full-siblings” and “parent-child”, which are distinguishable through framework (a). In this work, under framework (b), we show that extending the standard estimation of the autosomal coancestry to X-chromosomal transmission can be a powerful approach to bring into light distant genealogies and to discern specific pedigrees, mainly if combined with the estimation of the autosomal counterpart. Indeed, for the latter purpose, the combined statistics for autosomes and X-chromosomal markers can be as powerful as that obtained with the, so far, unbeatable framework (a).