Copy number variation in the horse genome

Abstract

We constructed a 400K WG tiling oligoarray for the horse and applied it for the discovery of copy number variations (CNVs) in 38 normal horses of 16 diverse breeds, and the Przewalski horse. Probes on the array represented 18,763 autosomal and X-linked genes, and intergenic, sub-telomeric and chrY sequences. We identified 258 CNV regions (CNVRs) across all autosomes, chrX and chrUn, but not in chrY. CNVs comprised 1.3% of the horse genome with chr12 being most enriched. American Miniature horses had the highest and American Quarter Horses the lowest number of CNVs in relation to Thoroughbred reference. The Przewalski horse was similar to native ponies and draft breeds. The majority of CNVRs involved genes, while 20% were located in intergenic regions. Similar to previous studies in horses and other mammals, molecular functions of CNV-associated genes were predominantly in sensory perception, immunity and reproduction. The findings were integrated with previous studies to generate a composite genome-wide dataset of 1476 CNVRs. Of these, 301 CNVRs were shared between studies, while 1174 were novel and require further validation. Integrated data revealed that to date, 41 out of over 400 breeds of the domestic horse have been analyzed for CNVs, of which 11 new breeds were added in this study. Finally, the composite CNV dataset was applied in a pilot study for the discovery of CNVs in 6 horses with XY disorders of sexual development. A homozygous deletion involving AKR1C gene cluster in chr29 in two affected horses was considered possibly causative because of the known role of AKR1C genes in testicular androgen synthesis and sexual development. While the findings improve and integrate the knowledge of CNVs in horses, they also show that for effective discovery of variants of biomedical importance, more breeds and individuals need to be analyzed using comparable methodological approaches.

Figure 1. A CNVR map of the horse genome.

Green line – loss; red line – gain; yellow line – complex; black dots – genes involved.

Figure 2. Chromosome-wise distribution of genic and intergenic CNVRs in the horse genome.

Figure 3. Gene Ontology classifications of copy number variable genes in horses.

Figure 4. A summary diagram of all CNV studies in horses and their contribution to the integrated CNV dataset.

Numbers in arrow-heads denote the contribution of each study to the common pool of 301 shared CNVRs; numbers in arrow-tails denote the total and novel (separated by colon) CNVRs per study; CGH, SNP and NGS denote the platforms used for CNV detection.

Figure 5. Validation of CNVRs by PCR.

A. gains and B. losses in Chr27 (37.3 Mb; probe Gs_27_37371896) involving CUB and Sushi multiple domains 1 (CSMD1) gene; C. Loss in Chr20 (24.8 Mb; probe Eic_20_24841849) involving olfactory receptors; n – number of individuals analyzed.

Figure 6. Chromosomal assignment and validation of a CNVR in chrUn (529–531 kb) by FISH.

A. Mapping the CNVR to chr19q12–q13 by FISH with BAC 132B13 (red); green - a control BAC with UMPS gene in chr19q21 ; B. The CNVR (red) in interphase chromosomes of a Percheron; green – a single-copy control probe; C. The CNVR (red) in interphase chromosomes of a Thoroughbred (Twilight). Note the difference in copy numbers between the Percheron and the Thoroughbred, as well as between homologous chromosomes.

Figure 7. Schematic of the homozygous deletion in chr29, 28.6–28.8 Mb in two XY DSD horses.

A. chr29 ideogram showing the location of AKR1C genes and a control gene CREM; B. Detailed map of the CNVR showing the location of genes (black horizontal bars) and CGH signal log2 values for 47 array probes in XY DSD and control horses; C. FISH results with a BAC 23N13 spanning the deletion (green signal) and a control BAC 76H13 for CREM from a non-CNVR (red signal); D. PCR with CNVR-specific primers in XY DSD and control horses.

Figure 8. Genetic relationships of horse breeds studied for CNVs.

A Maximum Likelihood tree showing genetic relationships of the horse breeds that have been studied for CNVs; * new breeds added in this study (except Swiss Warmblood);** breeds involved in 2 or more studies. Numbers denote bootstrap values.