A genome‐wide association study of morphometric traits in dromedaries

Abstract Background Investigating genomic regions associated with morphometric traits in camels is valuable, because it allows a better understanding of adaptive and productive features to implement a sustainable management and a customised breeding program for dromedaries. Objectives With a genome‐wide association study (GWAS) including 96 Iranian dromedaries phenotyped for 12 morphometric traits and genotyped‐by‐sequencing (GBS) with 14,522 SNPs, we aimed at identifying associated candidate genes. Methods The association between SNPs and morphometric traits was investigated using a linear mixed model with principal component analysis (PCA) and kinship matrix. Results With this approach, we detected 59 SNPs located in 37 candidate genes potentially associated to morphometric traits in dromedaries. The top associated SNPs were related to pin width, whither to pin length, height at whither, muzzle girth, and tail length. Interestingly, the results highlight the association between whither height, muzzle circumference, tail length, whither to pin length. The identified candidate genes were associated with growth, body size, and immune system in other species. Conclusions We identified three key hub genes in the gene network analysis including ACTB, SOCS1 and ARFGEF1. In the central position of gene network, ACTB was detected as the most important gene related to muscle function. With this initial GWAS using GBS on dromedary camels for morphometric traits, we show that this SNP panel can be effective for genetic evaluation of growth in dromedaries. However, we suggest a higher‐density SNP array may greatly improve the reliability of the results.

GWAS using GBS on dromedary camels for morphometric traits, we show that this SNP panel can be effective for genetic evaluation of growth in dromedaries. However, we suggest a higher-density SNP array may greatly improve the reliability of the results.

K E Y W O R D S
Camelus dromedarius, candidate genes, genotyping by sequencing, single nucleotide polymorphism (SNP) INTRODUCTION Since ancient times, domestic even-toed ungulates have played pivotal roles for man, being exploited for meat and milk, for fibre production, as beasts of burden for transport in agricultural/ rural oriented community; they were even worshipped (Barazandeh et al., 2019). These animals have served for traditional technologies since the very early era of domestication (Barazandeh et al., 2016). The Bactrian camel was domesticated, probably in northern of Iran, northeast Afghanistan or southwestern Turkmenistan (Mohammadabadi et al., 2021). Camels have been producing meat, milk and wool in desert conditions for thousands of years. Among the 35 million camels in the world, 95% are one-humped camels (Hashim et al., 2015). The number of Iranian camels is 152,346, which are being bred in the southern and central deserts of Iran (FAO, 2021). The most breeding units are in Yazd, Kerman, Semnan and Sistan, and Baluchistan provinces. Camel meat in Iran, with the production of four and a half thousand tons, includes 1% of the country's total red meat production (Khodai, 2004). This is even though the amount of meat and milk production in Iran is very low in general, and the declining trend of camel population in the country in recent years is worrying. To foster camel breeding for sustainable animal protein production from desert climate resources, attention to genetic improvement is of particular importance. Successful breeding programs in camels face basic problems such as lack of sufficient phenotype records, missing pedigree, small herd size and lack of kinship relationships. Furthermore, the morphometric recording can be notoriously difficult and prone to injuries caused by dromedaries not accustomed to measuring procedures. Thus, no systematic breeding program has been carried out in Iran, so far. In camel breeding programs, survival traits, reproductive traits (pregnancy rate, calving interval and age at first calving) and production traits including calves' and camel's weights should be considered as breeding goals (Vatankhah et al., 2019). Vatankhah et al. (2019) reported the economic value of calving interval to be −13.19$/month. To preserve and improve the genetic stock of camels and increase the productivity of meat production, it seems necessary to use new genomic technologies.
Due to the long generation interval of camels, it takes many years to achieve genetic improvement using traditional breeding, and it is practically impossible to achieve this with the usual methods. On the other hand, the cost of breeding in the traditional way and based on progeny testing is very high compared to modern genomic methods.
By using modern breeding methods, including the use of customised low-density SNP arrays, the speed of genetic progress can be significantly increased by using selection at a young age. The SNPs and genes associated with morphometric traits, especially height at withers, body length, hip width, breast circumference and scrotal circumference in cattle have been presented with several GWAS studies Cole et al., 2011;Terakado et al., 2018). The heritability of these traits is mostly reported to be moderate to high Munim et al., 2013;Zhang et al., 2017). Despite the fact that the genomic structure of cattle morphometric traits is highly polygenic, there are many similarities with other livestock species, humans and mammals in general (Bouwman et al., 2018;Pryce et al., 2011). The identified regions were mainly involved in biological functions such as regulation of embryonic development, skeletal development, regulation of cell cycle or cell division, homeostasis and lipid metabolism (Bouwman et al., 2018;Pryce et al., 2011;Setoguchi et al., 2011). More interestingly, different studies have identified significant overlaps in genomic architecture and genomic relationships between morphometric traits and cattle performance characteristics, including body weight, carcass traits, feed consumption, reproduction, and health, and in other words, phenotypic correlation between morphometric traits and animal performance such as body weight, milk production is similar in different breeds Bilal et al., 2016;Lukuyu et al., 2016;Misganaw et al., 2013;Pryce et al., 2011;Thomas et al., 2021). In addition, models have been developed to predict body weight using chest girth, body length, or height at whither (Lukuyu et al., 2016;Thomas et al., 2021), which suggests that morphometric traits are important predictors of animal performance in African domestic livestock herds, which have a weak recording system. Addressing the principles of selection, several studies showed that population morphometric diversity is the result of selection for adaptive and sociocultural interests in small livestock production in Africa (Kabi et al., 2015;Yougbaré et al., 2020;Yougbaré et al., 2021). Therefore, investigating the genomic regions associated with morphometric traits in camels is valuable, because it allows a better understanding of adaptive and productive features. In addition, it opens the prospect of effective use of morphometric traits in recording the basic phenotype for any potential community-based breeding program . With a GWAS approach, we aimed at identifying candidate genes potentially associated to morphometric traits to develop the basis for a sustainable breeding program for dromedaries in Iran.

SIMPLE SUMMARY
With a genome-wide association study (GWAS) including 96 Iranian dromedaries phenotyped for 12 morphometric traits and genotyped-by-sequencing (GBS) with 14,522 Single Nucleotide Polymorphisms (SNPs), we aimed at identifying associated candidate genes. We detected 59 SNPs located in 37 candidate genes potentially associated to morphometric traits in dromedaries. The top associated SNPs were related to pin width, whither to pin length, height at whither, muzzle girth, and tail length. We identified three key hub genes in the gene network analysis including Actin

Phenotypic measured traits
The number of 255 records of 12 morphometric traits were collected from 96 Iranian male camels during birth to 6 months old (

GBS technique and Bioinformatic analysis
The 96 Blood samples were gathered from jugular vein using EDTA tubes. The genomic DNA was extracted using the modified saltingout method and quantified using spectrophotometry and checked for quality on a 1% agarose gel (Javanrouh et al., 2006). The samples were genotyped using GBS, which involves several steps; first, cutting the DNA and producing fragments using two restriction enzymes (EcoR1 and HinF1); second, connecting the adapter to the cut fragments and removing very small fragments; and finally amplification of fragments using DNA polymerase. Paired-end (150 bp) next-generation sequencing (10 X) on the Illumina HiSeq 2000 platform was performed by Bayan Gene Pars Company. The produced sequences were mapped with the assembled genome of the camel (assembly accession: GCA_000803125.3) using the BWA-MEM algorithm and software (BWA) (Li & Durbin, 2010). PCR Duplicates were detected using the Picard tool and disregarded in downstream analyses both by GATK  and SAMtools (Pembleton et al., 2013).
SNPs were called across the GBS data using GATK and with a filter for a minor allele frequency (MAF < 0.01). The association between SNPs and morphometric traits was investigated using a linear mixed model with PCA and kinship matrix by TASSEL software (Bradbury et al., 2007). Geographical region and calf age were included as fixed and covariate effects in the model, respectively. The MLM_PCA+K statistical model was applied as follows: where y was the phenotype value; α was the vector of SNP effects; β was vector of population structure effects based on PCA; u was vector TA B L E 1 Mean, standard error (SE) and coefficient of variation (CV) of 255 morphometric records in dromedaries. The suggestive significant Bonferroni p-value thresholds were set (−log p-value > 4) using the GEC software tool (Li et al., 2012). The Interacting Genes (STRING) version 11.5 (https://string-db.org) was used to explore protein-protein interaction (PPI) networks (Shannon et al., 2003). A combined interaction score of > 0.4 was considered significant. The PPI networks were visualised using Cytoscape software (http://www.cytoscape.org) (Szklarczyk et al., 2015). The p-value ≤ 0.05 was considered to indicate a statistically significant difference. The topology scores of the nodes in the PPI network were estimated based on closeness centrality, betweenness centrality and degree centrality

Morphometric measurements
Descriptive statistics of morphometric traits are given in Table 1, while the corresponding correlation coefficients are visualised in Figure 2.
Coefficient of variation (CV) of morphometric records ranged 0.10 to 0.27. There was high correlation among morphometric records so that most of them were significantly correlated.

F I G U R E 2
The correlation coefficients of morphometric records in dromedaries.

Associated SNPs and candidate genes
There were no determined clusters among geographical regions.
Only 1.6% and 1.4% of the genetic variance was illustrated with PC1 and PC2, respectively, which demonstrated that calves are relatively homogeneous. A genome-wide association study (GWAS) was performed using 14,522 markers and detected 59 SNPs associated with morphometric traits (Figure 3 and Table 2) at a cut-off −log p-value > 4. The most associated SNPs were related to pin width and whither to pin length (

Protein-protein interaction (PPI) network construction and hub gene identification
A PPI network was established to investigate the interaction between the common DEGs using the STRING database and Cytoscape software ( Figure 4). Further analysis revealed that these candidate genes were linked to nine ontological processes including neuron differentiation, mesoderm formation, palate development, cytoplasmic ribonucleoprotein granule, nucleoplasm, Golgi apparatus, cytosol, nuclear matrix, transcriptional repressor activity, RNA polymerase II transcription regulatory region sequence-specific binding (Table 3).

TA B L E 2
Genome and chromosome-wide significant SNPs and potential candidate genes associated with morphometric traits in dromedaries.

Associated SNPs
We produced over 14k markers using genotyping-by-sequencing (GBS) technique to carry out a genome-wide association study (GWAS) for investigating the morphometric traits of 96 dromedaries. A total of 59 SNPs were associated with morphological measures such as height at whither, muzzle circumference, pin width, tail length, and whither to pin length, with the top significant SNPs located in Chr 11, 14 and 18.

F I G U R E 4 Network of candidate genes related to morphometric traits in dromedaries.
(2022) identified the seven SNPs strongly associated with stature in cattle using Genome-wide association studies.

Candidate genes and PPI network construction
We investigated the functions of the top associated candidate genes.
ACTB gene codes the cardiac muscle actin, the skeletal muscle actin and the cytoplasmic beta-actin (Czosnek et al., 1983). DDX27 and DEAD-Box RNA helicase are required for skeletal muscle growth and regeneration (Bennett et al., 2018). ZNF326 regulates cell growth (Yu et al., 2019). Jiang et al. (2021) reported that the ZNF521 gene is associated with body size in cross-bred sheep. TBX15 is required for skeletal development (Yu et al., 2019). Eight prioritised regulatory SNPs in the TBX15/WARS2 region are risk candidates for obesity and/or osteoporosis risk (Zhang et al., 2020). DEXI and CIITA genes have effects on the immune system (Edgar et al., 2001;Mehta et al., 2011). CLEC16A and ACTR3B were associated with hip width in Simental and Angus cattle (Edgar et al., 2001). Ramírez-Ayala et al. (2021) investigated that PTPN1 gene can be effective in thermogenesis and hair development.
Also the SNP rs3787348 in PTPN1 was associated with the body mass TA B L E 3 Candidate genes significantly enriched to morphometric traits-related GO terms and the KEGG pathway.  (Bera et al., 2008). Skeletal muscle growth is regulated by DDX27 gene through modulation of translational processes (Bennett et al., 2018). Thomas et al. (2021) reported that haploinsufficiency of ARFGEF1 gene can cause developmental delay (Teumer et al., 2016).

Genes
Also, results showed SUPT20H gene effects on rheumatoid arthritis (Veyssiere et al., 2019). SNAI1 knockdown decreased the expression of genes related to cytoskeleton rearrangement and extracellular matrix (ECM) remodelling (Sun et al., 2018). Teumer et al. (2016) confirmed that TNS3 gene associated with circulating IGF-I and IGFBP-3 concentrations. The PGIS gene is effective in cytoskeleton differentiation (Murray et al., 2018 and cytokine activity) that were associated with growth (Murray et al., 2018). In this research, the lack of morphometric data and pedigrees, small herd size and missing connectedness, and genetic assessments are the main limitations.

CONFLICT OF INTEREST STATEMENT
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

ETHICS STATEMENT
All

DATA AVAILABILITY STATEMENT
The data sets generated and during the current study are available in