Whole genome sequence data of 68 sheep representing 43 breeds from 19 countries was generated by the International Sheep Genomics Consortium, a partnership of researchers and funding agencies (http://www.sheephapmap.org/). The SAMtools suite (http://samtools. sourceforge.net) was used to extract reads aligned to genome regions (O. aries genome sequence NC_019458.2) that correspond to ovine MX1 (Gene ID: 443146) and MX2 (Gene ID: 780441) sequences. The reads obtained were re-aligned to MX1 or MX2 sequences using Geneious v. 10.1.3 (Biomatters, Auckland, New Zealand; 69Kearse et al. 2012) and variant calling was undertaken. Potential variants were further analysed to produce lists of variant sequences that fulfil the following criteria: (i) the nucleotide variation was observed in regions of 10-fold or higher read coverage in each individual dataset; (ii) the observed nucleotide variation with a minor allele read frequency was greater than 0.45 for at least one individual; (iii) homozygotes all had one of the two possible nucleotides (or gaps) observed in the heterozygotes, or were all homozygous for the non-reference allele; and (iv) the observed nucleotide variation was of the same type in at least two heterozygous individuals and was the same type in all heterozygous individuals considered. To ensure that the frequency of variants was not overly affected by sample size, a “suitable coverage region” criterion was added. Suitable coverage regions were defined as the part of the alignment where the total number of reads across all individuals with a minimum read coverage of 10 exceeded 99 reads.