Methods

Sampling and DNA sample collection
Nasal samples were collected using Performagene LIVESTOCK's nasal swab DNA collection kit (DNA Genotek Inc., Kanata, ON, Canada) from 5 cattle populations-Ambo, Borana, Arsi, Horro, and Danakil-inhabiting different agro-ecologies. Sampling was carried out by considering potential geo-environmental gradients (highland, low-lands), production systems (mixed-crop livestock, pastoral/agro-pastoral), and ethnic groups raising predominantly the populations (Table 1, Fig. 1). Representative sites were selected based upon secondary data obtained from livestock departments and expertise. Herdsmen and owners of the animals were contacted to ensure that the animals were not genetically related. Animals were randomly sampled from multiple herds. Genomic DNA was extracted from nasal samples based on laboratory protocols. In addition, DNA samples from Hanwoo cattle were included for comparison purposes.

Breeds/populations descriptions
The Danakil cattle breed is grouped under Sanga and originated in East and Northeast Africa, where Sanga cattle evolved. The Horro breed was formed as a result of interbreeding between zebu-Sanga and Sanga-zebu admixture populations. The large-sized Borana breed descended from secondary cattle domestication in the arid areas of the 'Fertile Crescent' about 5000 BP [7]. The breed is predominantly distributed in arid and semi-arid agro-ecologies of the Borana plateau. The small-sized Arsi cattle probably developed from a group of small short-horned Abyssinian zebu by the highland Oromo people. The Ambo cattle is grouped under Small East African zebu and is believed to have descended from the recent introductions of zebu into Africa [8]. Pictures of sampled animals are depicted in Fig. 2.

Genotyping and quality control
In total, 192 animals representing 5 cattle populations of Ethiopia were genotyped for 8,773 SNPs with the Illumina Bovine 8K SNP BeadChip (Illumina, San Diego, CA, USA) according to Illumina protocols [9]. Over the loci, the average GenTrain score, 10% GenCall (10% GC) score, and 50% GenCall (50% GC) score were 0.86, 0.87, and 0.83, respectively. In this experiment, about 99.91% of the markers identified had a GenTrain score greater than the minimum acceptable value (0.25) (http://www.illumina.com). All of the samples had at least a 0.20 or greater 10% GC score, while the 50% GC score ranged from 0.75 to 0.91, with the average call rate of 0.913. For Hanwoo cattle, about 51,162 SNPs were genotyped using the Illumina Bovine 50K SNP BeadChip [10]. 8K and 50K genotypes were merged using Golden Helix SNP Variation Suite software version 7 (Golden Helix, Inc., Bozeman, MT, USA). To ensure the highest quality of data from the 8K and 50K SNP BeadChips, duplicate marker genotypes were identified and dropped. Finally about 7,045 SNPs that were common for both Bead-Chips were subjected to analysis. Markers selected for diversity analysis were all located on autosomal chromosomesand had at least an 80% call rate.

Data analysis
Frequency of minor allele frequency (MAF), distribution, and deviation from Hardy-Weinberg equilibrium (HWE) were estimated using Golden Helix SNP Variation Suite version 7 (Golden Helix, Inc.). The SAS GLM package version 9.1 (SAS Institute Inc., Cary, NC, USA) procedure was used to test the effect of breed on MAF variation. Observed and expected heterozygosities and hierarchical F-statistics were estimated using GENALEX software [11].