Genome-wide association study reveals novel candidate genes for litter size in Markhoz goats

Abstract

Introduction: The Markhoz goat is the only breed that can produce high-quality fiber called mohair in Iran; however, the size of its population has faced a dramatic decline during the last decades, mainly due to the reluctance of farmers to rear Markhoz goats caused by a reduction in goat production income. Litter size at birth (LSB) and weaning (LSW) are two economically important reproductive traits for local goat breeders and have the potential of increasing the population growth rate. The present study was aimed to identify possible genomic regions that are associated with LSB and LSW in Markhoz goats using a genome-wide association study (GWAS).

Methods: To this end, 136 Markhoz goats with record(s) of kidding were selected for GWAS using the Illumina Caprine 50K bead chip. The individual breeding values (BV) of available LSB and LSW records estimated under an animal mixed model were used as the dependent variable in the GWAS, thereby incorporating repeated categorical variables of litter size.

Results: Four SNPs on chromosomes 2, 20 and 21 were identified to be significantly associated (FDR p < 0.05) with LSB after multiple testing correction under a Bayesian-information and Linkage-disequilibrium Iteratively Nested Keyway (BLINK) model. Least-square analysis was performed to investigate the effects of detected genotypes on LSB. Ultimately, the GWAS results introduced six candidate genes, including GABRA5, AKAP13, SV2B, PPP1R1C, SSFA2 and TRNAS-GCU in a 100 kb adjacent region of the identified SNPs. Previous studies proposed functional roles of GABRA5 and AKAP13 genes in reproductive processes; however, the role of other candidate genes in reproduction is not clear.

Conclusion: These findings warrant further investigation for use in marker-assisted selection programs in Markhoz goats.

Keywords: Markhoz goat; genome-wide association study; litter size; prolificacy; reproduction.

Figure 1

The distribution of litter size at birth (A); litter size at weaning (B); PBVs for litter size at birth (C); and PBVs for litter size at weaning (D).

Figure 2

The Manhattan and quantile–quantile (Q–Q) plots of -log 10 (p-value) for litter size at birth (LSB) (A); and litter size at weaning (LSW) (B) in the Markhoz goats using the BLINK GWAS model. The green horizontal line indicates the FDR adjusted p-value threshold of 0.05.

Figure 3

The Q–Q plot for litter size at birth (LSB) fitting Bayesian-information and Linkage-disequilibrium iteratively nested keyway (BLINK) (A); general linear model (GLM) (B); mixed linear model (MLM) and multi locus mixed model (MLMM) (C); compressed mixed linear model (CMLM) (D); enriched compressed mixed linear model (ECMLM) (E), factored spectrally transformed linear mixed model (Fast-LMM) (F); settlement of MLM under progressively exclusive relationship (SUPER) (G); fixed and random model circulating probability unification (FarmCPU) (H); and efficient mixed-model association (EMMA) and efficient mixed model association eXpedited (EMMAX) (I) GWAS models.