Statistical confidence for likelihood-based paternity inference in natural populations

摘要: Paternity inference using highly polymorphic codominant markers is becoming common in the study of natural populations. However, multiple males are often found to be genetically compatible with each offspring tested, even when probability excluding an unrelated male high. While various methods exist for evaluating likelihood paternity nonexcluded male, interpreting these likelihoods has hitherto been difficult, and no method takes account incomplete sampling error-prone genetic data typical large-scale studies systems. We derive ratios taking typing error, define a statistic delta resolving paternity. Using allele frequencies from population question, simulation program generates criteria that permit assignment most likely known level statistical confidence. The number candidate males, proportion sampled gaps errors data. explore potentially confounding effect relatives show robust their presence under commonly encountered conditions. demonstrated intensively studied red deer (Cervus elaphus) on island Rum, Scotland. Windows-based computer program, CERVUS, described this available authors. CERVUS can used calculate frequencies, run simulations perform parentage analysis all types markers.