Genomic scan for detection of copy number variations in Caspian and Turkman horses

Introduction: Horses have played an important role in the history of Iranians during different centuries. They kept horses for various aims such as agriculture, transportation, sport, food sources. Iran has a suitable climatic, social and economic potential for keeping and breeding horses, that is why it has been created different breeds using the selection and breeding. But due to problems such as mechanization, lack of government support, export ban, high costs of breeding and maintenance, import of foreign horses, lack of proper planning, interest in keeping horses has decreased. So, unfortunately, only a few native Iranian breeds remain. Additional investigation of the equine genomic architecture is critical for a better understanding of the equine genome, as well as for expanded comparisons across diverse mammalian species. Turkmen and Caspian horses are well-known breeds of Iranian horse breeds. These breeds were historically selected to perform distinct tasks and therefore may harbor a wealth of unique variation at the genome level. Copy number variation (CNV) along with single nonucleotide polymorphisms (SNPs) play a key role in genetic diversity in livestock species. CNVs, a term that refers to a change in the number of copies of a genomic segment, are responsible for more sequence differences between individuals than SNPs and are considered to be a major source of genetic variation contributing to differences in phenotypes (Beckmann et al, 2007). Several studies identified copy number variations in horses using different techniques (Doan et al, 2012). Part of these studies tried to establish associations between CNVs and a specific trait, a disease or even gene expression (Schurink et al, 2017). Most of these studies found either no association or inconclusive associations as the number of horses with phenotypic information or with specific CNVs were limited. For example, a 62 kb duplication on Equus caballus (ECA) chromosome 10 seemed to be related to recurrent laryngeal neuropathy (Dupuis et al, 2013). However, little is known about CNV in Iranian horses.
Materials and Methods: In this study, detection of CNVs and CNVRs were performed based on SNP data from Caspian and Turkman horse breeds were genotyped via Equine70k SNP beadchip. PennCNV software was only used to detect CNV on autosomes. The PennCNV algorithm was only applied to autosomes (command: -lastchr 26) to identify individual-based CNVs. To increase the confidence of the detected CNVs, quality control was performed by employing standard exclusions of the LRR (standard deviation of LRR) <0.3, a BAF drift <0.01 and a waviness factor <0.05. We classified the status of these CNV into two categories: “loss” (CNV containing a deletion) and “gain” (CNV containing a duplication). The CNVRs were determined by aggregating the overlapping CNVs with CNVRuler. BioMart in the Ensembl database and DAVID was employed to identify genes located in CNVRs and GO terms and KEGG pathway analyses respectively. Quantitative real-time PCR (qPCR) was applied to validate the CNVRs that were detected in this study.
Results and Discussion: A total of 202 and 105 CNVs and CNVRs were identified in the studied horses, respectively, which cover 1.08% of the horse genome. The number CNVs in Caspian breed were 1.6 times more than Turkmen. Also, the average size of CNVs in Caspian breed was longer than Turkmen. In both breeds, the genetic event of gain was higher than the genetic event of deletion. In Caspian breed, chromosomes 1, 3, 12, and in the Turkmen breed, chromosomes 1, 6 and 12 showed the most changes in CNVs, respectively. Functional analysis showed that the identified CNVRs overlapped with 434 genes and the most of these genes were common between the two horse breeds (more than 60%). Among these genes, PPARG and GALR have potential related with breed-specific traits. The KEGG pathway analysis also identified several pathways that are significantly enriched in olfactory sensory perception, chemical stimulus sensory perception, antigen processing, and G protein signaling pathway. Also, 60% of successfully detected CNVRs were confirmed by Real-time qPCR. The results of this study were compared with the results of eight other studies. For example, we concluded that the average size of the CNVRs detected by the 70k arrays and the 50K arrays were significantly larger than obtained by the CGH and NGS arrays. This may be due to the relatively low coverage and non-uniform distribution of SNP in the equine genome in SNP arrays. Possible reasons for the differences between our results and some CNV studies can be related to different parameters such as sample size and genetic background, different detection platforms and CNV retrieval algorithms, CNV definitions and CNVRs, as well as random error estimation methods (Pinto et al. 2011).
Conclusions: CNVs can describe part of the phenotypic diversity and adaptation evidence in Iranian horses. With regard to Genes identified in a number of cellular components, biological processes and molecular functions within CNVRs, the importance of such CNVRs and the possible effect needs to be studied and may interest insight into the functional and adaptive consequence of CNVs in horse. In total, the number of CNVs in the Caspian breed was greater than in the Turkmen breed, and also the CNV length in the number of copies in the Caspian breed was greater than in the Turkmen breed. In both breeds, there were overlapping genes with CNVRs that were significantly enriched in biological pathways, including sensory perception, immunity, and metabolism. This is the first CNV report on Turkmen and Caspian horses and the findings of this study could provide valuable information for better understanding of the horse genome and also the important performance traits with CNVRs and associated genes for the future studies in horse breeds.