Hiphop: Improved paternity assignment among close relatives using a simple exclusion method for biallelic markers. (3rd April 2021)

• Record Type: Journal Article
• Authors: Cockburn, Andrew; Peñalba, Joshua V.; Jaccoud, Damian; Kilian, Andrzej; Brouwer, Lyanne; Double, Michael C.; Margraf, Nicolas; Osmond, Helen L.; Kruuk, Loeske E. B.; van de Pol, Martijn
• Is Part Of: Molecular ecology resources. Volume 21:Number 6(2021)
• Page Start: 1850
• Page End: 1865
• Publication Date: 2021-04-03
• Subjects: Avian mating systems -- Cooperative breeding -- Fairy‐wren -- Malurus cyaneus -- Molecular sexing -- Parentage analysis; Molecular ecology -- Periodicals; 572.8
• DOI: 10.1111/1755-0998.13389
• Languages: English
• ISSNs: 1755-098X

Abstract

Abstract: Assignment of parentage with molecular markers is most difficult when the true parents have close relatives in the adult population. Here, we present an efficient solution to that problem by extending simple exclusion approaches to parentage analysis with single nucleotide polymorphic markers (SNPs). We augmented the previously published homozygote opposite test ( hot ), which counts mismatches due to the offspring and candidate parent having different homozygous genotypes, with an additional test. In this case, parents homozygous for the same SNP are incompatible with heterozygous offspring (i.e., "H omozygous I dentical P arents, H eterozygous O ffspring are P recluded": hiphop ). We tested this approach in a cooperatively breeding bird, the superb fairy‐wren, Malurus cyaneus, where rates of extra‐pair paternity are exceptionally high, and where paternity assignment is challenging because breeding males typically have first‐order adult relatives in their neighbourhood. Combining the tests and conditioning on the maternal genotype with a set of 1376 autosomal SNPs always allowed us to distinguish a single most likely sire from his relatives, and also to identify cases where the true sire must have been unsampled. In contrast, if just the hot test was used, we failed to identify a single most‐likely sire in 2.5% of cases. Resampling enabled us to create guidelines for the number of SNPs required when first‐order relatives coexist in the mating pool. Our method, implemented in the R package hiphop, therefore provides unambiguous parentage assignments even in systems with complex social organisation. We also identified a suite of Z‐ and W‐linked SNPs that always identified sex correctly.