Increased assignment rates of parentage assignment obtained by resampling populations

Parentage analysis provides indispensable information for the study of ecology and evolution. However, many parentage algorithms suffer from low assignment rates, especially when the number of molecular markers is less than adequate, or when prior knowledge about the study population is limited. The present study developed a novel pipeline, which uses resampling methods as an external extension, to achieve higher assignment rates of parentage analysis. The R package provided by this study helps CERVUS, one of the commonly used parentage software, to find more potential father–mother–offspring triads. Simulations showed that the idea of resampling offers great potential to find father–mother–offspring triads that are overlooked by CERVUS, while keeps the same level of statistical confidence. Performances were also tested based on simulated data. When eight polymorphic microsatellite markers are genotyped in a population of 200 individuals, for example, the pipeline improved the assignment rates by over 10%, while the error rates did not rise. This study highlights the importance of incorporating resampling methods into parentage algorithms. The framework of resampling methods is applicable to other parentage software to find more potential parent–offspring relationships.