Pride and prejudice : extra-group paternity in lions and the effects of marker density on kinship and relatedness estimates

Abstract: The development effective high-throughput sequencing techniques has allowed further insight to the mating systems of wild animal populations and challenged our perception that reproductive behaviour can be predicted from social structure. A recent study of lions (Panthera leo) in Etosha National Park, Namibia, found evidence of extra-group paternity (EGP) which has never before been confirmed among lions. Relatedness estimated from earlier research on lions in Selous Game Reserve, Tanzania, suggests that this population also deviates from a strict within-pride mating system. In this study, the same population of wild lions from Selous Game Reserve was visited two times, with twelve years in between, and with the aid of RAD-sequenced SNP markers, pedigrees were reconstructed and pairwise relatedness calculated to explore their mating system and compare with their social structure. Additionally, the effect of marker density (i.e. number of markers) on the precision of the kinship and relatedness estimates was explored. Evidence of EGP was found in the Selous population; of the paternal relationships detected 60% (n=15), were between members of different prides. This result calls for more studies of the reproductive behaviours of the different lion populations that remain to see how the behaviour is distributed. Increasing marker density positively affected the accuracy of the pairwise relatedness estimates in the range of 100-1000 SNPs. In the context of parentage analysis increasing the number of markers lead to relatedness values of the relationships detected to have a mean closer to the theoretically expected value of 0.5, however the number of relationships detected decreased considerably with marker densities above 200. Thus, future studies conducting parentage analyses should not simply use as many markers as possible, but instead consider the optimal number markers in the context of their data.