INTRODUCTION
The camel is an important livestock species in many regions of the world, and is used for meat, milk, skin production, as well as a means of transport, entertainment, racing competitions, dancing and beauty shows. It is adapted to extremely harsh environments [
1], and is well placed to cope with climate change. While providing financial benefit, often to economically-disadvantaged farmers, the camel has a substantial cultural value within rural communities. There are three species of camel, the dromedary (
Camelus dromedarius) from south Asia, the Middle East and North/East Africa, the Bactrian camel (
Camelus bactrianus) from central Asia, and wild Bactrian camel (
Camelus ferus) from China and Mongolia [
2]. However, the dromedary is by far the most common of the species and is the focus of the current paper.
Unlike other livestock species, little sy stematic breed improvement has been undertaken on the camel. Further, the genetic diversity has also not been explored thoroughly like other livestock species, and more work on this area has been conducted on the Bactrian camel than the dromedary [
3,
4]. The camel has a substantial potential for genetic improvement, particularly for meat and milk products. Given the adaptations of camels to be productive in environments too harsh for other livestock species, and with the urgency of climate change, it is timely to allocate resources to undertake genetic improvement programs in the camel. Thus, it is vital to study the genetic diversity and conserve germplasm resources of the camel [
5,
6].
Pedigree-based and genomic selection have the potential to play a major role for genetic improvement in the camel, however, it is also important to identify genes and genomic regions associated with growth and production traits, for understanding the genetic architecture of the traits and improving accuracy of selection. Genome-wide association studies (GWAS) have become routine for discovering the genes and narrow genomic regions associated with traits of interest. However, to conduct a GWAS, in addition to phenotypic records, the availability of sufficiently dense genetic markers is a prerequisite. High-density single nucleotide polymorphism (SNP) arrays have become available for several species and made a substantial contribution to breeding, and mapping disease traits in domesticated animals.
Currently no SNP chip is available for the dromedary camel. However, genotyping-by-sequencing (GBS) can be used to genotype a large number of SNPs with a much lower cost as compared to developing a SNP chip. Recently we published a study on SNP discovery and genetic diversity in the dromedary camel using this GBS technique [
7]. Sequencing as compared with SNP chips can provide a range of additional insights into genetic architecture, including information about existence of SNPs, copy number variation, and insertions/deletions (indels). Due to reduction in sequencing cost, it is now feasible to sequence and genotype all individuals in a population and then perform GWAS. This has particular appeal for livestock improvement in non-traditional species such as the camel.
A small number of studies have been conducted on ge netic associations with traits in camels. Almutairi et al [
8] reported a quantitative genetic analysis of growth and milk production traits of a Saudi camel population based on phenotypic and pedigree data, reporting on heritability estimates and other genetic parameters. Further, several studies have investigated trait associations with specific candidate genes. For example, Afifi et al [
9] have investigated the association between the growth hormone gene and body weight in dromedary camels, while Almathen et al [
10] reported associations of the melanocortin 1 receptor (
MC1R) and agouti signalling protein (
ASIP) genes with coat colour in dromedaries. Guo et al [
11] published a GWAS on haematological traits in the Bactrian camel in China. Recently, Bitaraf Sani et al [
12] presented a GWAS for an Iranian dromedary population for birth weight and average growth. However, there is no study, that the authors are aware of, in the dromedary camel, that undertake a GWAS across a range of weight-for-age traits.
The current study was designed with the objectives i) to undertake a GWAS to identify SNPs associated with growth at specific camel ages, and ii) for key significant SNPs identified in the GWAS, to investigate in detail how their effects change over the life of a camel. The analyses were performed separately for two breeds of dromedary camels commonly found in Pakistan, namely Marecha and Lassi. This is the first GWAS undertaken on dromedary camels in this region, and builds upon previous research on these two breeds in relation to growth and genetic diversity [
7,
13].
DISCUSSION
The camel is an excellent source of meat production in harsh and drought conditions. Genetic improvement of growth traits of the camel is critical to achieve increased meat and production efficiency. Therefore, it is necessary to find major genes that are associated with growth for future breeding selection. In this study, we perform GWAS of growth traits at different ages of their life, and significant association of SNPs have been identified.
GWAS is an efficient method for searching candidate genes associated with growth. However, the accuracy of GWAS relies on the population structure, as well as the existence of genome-wide linkage disequilibrium [
18]. Domestic animals have a simpler genetic diversity and population structure as compared with human populations [
19]. In this study, all Marecha camel samples were collected from the same farm, but a principal coordinate analysis on the SNP data revealed greater genetic diversity in that breed, compared with that Lassi camels collected from multiple farms [
7]. While reduced genetic diversity may impact the ability to detect genetic associations, in this study more genetic associations were detected for Lassi compared to the genetically more diverse Marecha breed. This may be due to the genes that did show diversity in Lassi have associations with growth traits. However, this would need to be confirmed with a larger number of animals to fully evaluate the genetic architecture. The main purpose of this work was to find SNPs associated with several body weights traits, namely, birthweight, weaning weight, and weights at 1, 2, 4, and 6 years old. Growth is a complex trait, involving aspects of cell proliferation, differentiation, muscle, bone and tissue development, fat, metabolism and nutrient absorption [
20]. Likely, each physiological aspect is under its own specific genetic control, so a search for associations will always be a somewhat ‘blunt tool’.
At the most stringent level of testing, (
q<0.01), there were five SNP associations detected for Marecha (one at 2 yr, four at 4 yr) and three associations detected for Lassi camels (one for weaning, two at 2 yr), as shown in
Table 2. However, many more were detected at less stringent levels (
q<0.05), particularly for Marecha camels at 4 yr, although the limited number of Marecha in this age group may influence this result.
In terms of effect sizes, for birthweight, effects sizes (allelic substitution effects) for significant SNP associations ranged up to −4.43±1.02 kg (Marecha) and −5.06±1.00 kg (Lassi), i.e. up to ~10% of average birthweight. For weaning weight, effect sizes ranged up to 20.9±3.8 kg (Lassi), i.e. ~20% of mean weaning weight (No significant associations were detected for weaning weight in Marecha camels). Maximum effect size of significant SNPs associated with weight at specific ages tends to increase with increasing age: for Marecha: 144.3±25.0 kg at 2 yr, and −159.2±25.1 kg at 4 yr, and for Lassi: 50.6±10.4 kg at 1 yr, and 107.4±19.4 kg at 2 yr. One explanation for the relatively large number of associations at the older groups may be that the effect sizes in these older cohorts are relatively larger, increasing their power of detection. However, the small sample sizes means that these results need to be treated with caution.
Also, of note was the existence of age-specific associations. The overwhelming majority of associations were found to be age-specific. Similarly, a study by Hadjipavlou and Bishop [
21] in Scottish Blackface sheep reported quantitative trait loci to be associated over specific age ranges. Nonetheless, some SNP were associated with weights at more than one age in the current study, particularly for Lassi camels with several SNP associations at one and two years of age (
Table 3,
Figure 9). These SNPs that show strong association with growth can be used as candidate gene in breeding program, perhaps incorporating a panel of genes that might respond at different ages. Nevertheless, it should again be considered that the associations at older ages are less clear, due to the reduced sample sizes. The significant SNPs were within 142, out of which 128 were protein coding and several of them have interesting biological functions relevant to the growth and metabolism. Further confirmation of these association with a larger sample size may be warranted before delving into functions of these candidate genes.
The current work builds on the study by Bitaraf Sani et al [
12] who reported the results of a GWAS on 96 dromedaries of several breeds in Iran. When the results of the Bitaraf Sani et al [
12] paper and the current study are compared, it was noted that a number of SNP associations within a ±1 Mb window were in common, notably on chromosomes 7, 11, 18, 25, 31, and 33 (
Supplementary Table S2). For example, there were four neighbouring SNPs on chromosome 18 (~ 30.3 Mb) relating to Marecha body weight in the current study which were within 0.6 Mb of eight neighbouring SNPs in the Iranian study of SNPs associated with body weight for which the authors suggested candidate genes of dexamethasone-induced protein (
DEXI), testis-specific Y-encoded-like protein 4 (
TSPYL4) and Class II transactivator (
CIITA). The current study provides a GWAS for body weight at a range of specific ages, and also tracks the effects of selected SNPs across ages of the camel. The novelty of this work has been enhanced by using the output from the growth curve analysis as additional weight phenotypes to detect effects of SNPs at different ages of animals Nevertheless there are some limitations to the current study, one major issue being the relatively small number of animals in the study. While adequate for the genetic diversity and population structure reported in Sabahat et al [
7], this is not ideal for a GWAS. Given the genetic difference between the two breeds considered here, it was necessary to perform separate GWAS as Marecha and Lassi, and this would reduce the power of detecting associations compared with a combined analysis. Nevertheless, the work presented here can be considered as laying the foundations for a more comprehensive map. This will facilitate development of marker-assisted selection approaches, and with increase in camel numbers genotyped, the possibility of a genomic selection approach for breed improvement in camels could be considered, as is currently conducted in other livestock species.