A method to improve confidence in paternity assignment in an open mating system

Molecular techniques have allowed researchers studying mating systems to determine the identity of extra-pair sires, providing more accurate measures of individual realized reproductive success. Yet, an existing problem in such studies is the inability to assign paternity to individuals that have not been captured. This frequently arises when only a proportion of the population is sampled or when visitors from outside the study area have access to the breeding population. It is therefore difficult to assign paternity in situations where not all candidate sires are sampled because some assignments may be incorrect, especially when using a likelihood-based approach. This study outlines a method that combines two different programs, GERUD 1.0 and CERVUS 2.0, to increase confidence in paternity assignment. The benefit of using these programs in conjunction is that GERUD 1.0 can reconstruct genotypes of males that are not sampled in families where the female was sampled, and CERVUS 2.0 can use this information to better assign paternity because more information is provided. We show how applying this method to Least Flycatchers (Empidonax minimus), a sub-oscine bird with an open mating system, substantially increases confidence in paternity assignments.