Assessing probability of ancestry using simple sequence repeat profiles: applications to maize inbred lines and soybean varieties.

Determining parentage is a fundamental problem in biology and in applications such as identifying pedigrees. Difficulties inferring parentage derive from extensive inbreeding within the population, whether natural or planned; using an insufficient number of hypervariable loci; and from allele mismatches caused by mutation or by laboratory errors that generate false exclusions. Many studies of parentage have been limited to comparisons of small numbers of specific parent-progeny triplets. There have been few large-scale surveys of candidates in which there is no prior knowledge of parentage. We present an algorithm that determines the probability of parentage in circumstances where there is no prior knowledge of pedigree and that is robust in the face of missing data and mistyped data. The focus is parentage of an inbred line having uncertain ancestry. The algorithm is a variation of a previously published hybrid-focused algorithm. We describe the algorithm and demonstrate its performance in determining parentage of 43 inbred varieties of soybean that have been profiled using 236 SSR loci and from seven inbred varieties of maize that were profiled using 70 SSR loci. We include simulations of additional levels of missing and mistyped data to show the algorithm's utility and flexibility.

Assessing probability of ancestry using simple sequence repeat profiles: applications to maize hybrids and inbreds.

Determination of parentage is fundamental to the study of biology and to applications such as the identification of pedigrees. Limitations to studies of parentage have stemmed from the use of an insufficient number of hypervariable loci and mismatches of alleles that can be caused by mutation or by laboratory error and that can generate false exclusions. Furthermore, most studies of parentage have been limited to comparisons of small numbers of specific parent-progeny triplets thereby precluding large-scale surveys of candidates where there may be no prior knowledge of parentage. We present an algorithm that can determine probability of parentage in circumstances where there is no prior knowledge of pedigree and that is robust in the face of missing data or mistyped data. We present data from 54 maize hybrids and 586 maize inbreds that were profiled using 195 SSR loci including simulations of additional levels of missing and mistyped data to demonstrate the utility and flexibility of this algorithm.