An Illumina 50 K SNP array for collared flycatcher has recently been developed by selecting markers from >10 million SNPs identified in genomic resequencing of 10 unrelated collared flycatchers (from our study population) and 10 pied flycatchers Ficedula hypoleuca (Kawakami et al. 2014 ). The bulk of markers were chosen based on a number of criteria set to maximize the usefulness in collared flycatchers, including polymorphism level in the sequencing sample, even distribution across the genome as judged by comparative map information vis-à-vis the zebra finch linkage map and, if possible, inclusion of at least two SNPs from all scaffolds >25 kb in a preliminary genome assembly version. Five thousand markers on the array were selected to represent potentially fixed differences between the two sister species and were thus generally less informative for intraspecific analyses.
Genotyping was done with an Illumina iScan instrument. Markers that failed to pass the quality filtering for genotype calling were removed from subsequent analysis. Deviation from Hardy–Weinberg equilibrium (HWE) was tested for in the parental generation using plink version 1.07 (Purcell et al. 2007 (link)). After filtering out SNPs deviating from HWE, Mendelian inheritance was inspected for the remaining markers using GenotypeChecker (Paterson & Law 2011 (link)). In total, 38 900 markers were polymorphic in the pedigree, of which 37 443 segregated with a minor allele frequency (MAF) >0.05. Among these, there were 845 putative Z-linked markers. The low proportion of loci with rare alleles illustrates the value of selecting markers based on prior information of polymorphism levels, in this case from whole-genome resequencing, in the same population.
The inheritance analysis revealed 89 individuals with at least one marker that did not follow Mendelian patterns. As extra-pair paternity (EPP) is known to occur frequently in the collared flycatcher (Sheldon & Ellegren 1999 (link)), individuals with a high proportion of markers deviating from expected Mendelian segregation likely result from EPP. We therefore removed 46 individuals in which >100 markers showed inconsistent inheritance. The remaining 43 individuals (of the 89 individuals with >1 error) had 1–15 markers with Mendelian inconsistency and were retained; however, the inconsistent markers (181 in total) were removed from the subsequent analysis in all individuals. In the end, we used genotype data from 609 individuals and 37 262 markers for linkage analysis. The average number of informative meioses in the pedigree across all markers was 187.