SNP associations and genetic-parameter estimation for nose and hoof pigmentation in Corriedale sheep

Animals

The data analysed were collected as part of the project initiated by the Performance Corriedale Group to improve this breed (Blackwood et al. 2021). Corriedale lambs (n = 764) were produced from three rounds of artificial insemination and one natural mating from April 2017 to 2019 (birth group). The matings resulted in 1–60 progeny per sire (n = 44). Sex, number of lambs born per ewe (type of birth), number of lambs reared per ewe (type of rearing), and slaughter date were recorded. Animals were kept on a commercial property in Evandale, Tasmania. Animal ethics approval was not required as measurements were normal management operations and, therefore, exempted.

Measures

Visual scores, wool-production, wool-quality, meat-production and meat-quality traits were measured as described in Table 1. Visual scores included nose pigmentation (NOSEP), hoof pigmentation (HVSP), face cover (FC), back conformation (BACK), and body wrinkle score (BDWR). All visual scores were based on a 1–5 scoring system (Australian Wool Innovation and Meat and Livestock Australia 2013) and were scored by two or three scorers. The mean of these scores was used in the analysis. Pre-slaughter measures were weaning weight at ~5 months of age (WWT), post-weaning weight (PWWT) at ~6–7 months of age, scanned eye-muscle depth (PEMD), scanned fat depth (PFat) and height (Ht). PEMD and PFat were measured using an ultrasound scanner. Ht was measured from the ground to withers. Post-slaughter measures included eye-muscle depth (EMD), eye-muscle width (EMW), eye-muscle area (EMA), hot standard carcass weight (HSCW), fat depth (GRFat), shear force of Longissimus lumborum (SF5) and intramuscular-fat percentage (IMF). SF5 was measured, with six replications, by using a Lloyd texture analyser fitted with a Warner–Bratzler head. Intramuscular fat (IMF) was measured in Longissimus lumborum via chemical extraction. Wool-quality traits included greasy fleece weight (GFW), percentage of fibres under 30 microns (CF), wool-staple length (SL), staple-structure score (SSTRC), wool colour score (COL) and wool character score (CHAR).

Genomic relationship matrix

The 764 lambs were genotyped using the Illumina OvineSNP50 BeadChip. The SNPs with a minor allele frequency of <0.01, and on the X and Y chromosome, were removed. The SNP data were used to construct a genomic relationship matrix (GRM) by using VanRaden’s first method (VanRaden 2008).

Heritabilities, genetic and phenotypic correlations

Univariate and bivariate analyses were conducted using ASReml-R (ASReml, version 4.1.0; Butler et al. 2018; https://asreml.kb.vsni.co.uk/) in R (version 4.2.1). One of three models was fitted to the traits (Table 2) and the levels of fixed effects are described in Table 3. The GRM was fitted as a random effect in each model. Sex was recorded directly from the lambs or, if it was not recorded, it was inferred using genotype information. For carcass traits, contemporary group was defined as a function of birth group and kill date. Specifically, one cohort (2019A) was slaughtered on two different days. Heterozygosity for each sample was estimated from the genotype data as the average heterozygosity across all SNPs.

Genome-wide association study

A genome-wide association study (GWAS) was conducted to identify regions associated with nose and hoof pigmentation. A mixed model was used to test association of pigmentation scores with the alleles of SNP genotype. All SNPs across the genome (Oar_v3.1) were tested with one SNP in the model at a time in the model. The model was fitted using GCTA software (Yang et al. 2011), by using the following model:

In this model, y represents the quantitative score, a is the mean term, c represents effect of fixed effect z, b is the additive effect of a SNP tested for association, x denotes the genotype indicator variable coded as 0, 1 or 2, g is the polygenic effect (random effect) captured by the GRM, calculated using all autosomal SNPs, and e is the residual. The fixed effects were as described in the section above.

The P-values and effect size for all SNPs across the genome were estimated. The genome-wide false-discovery rate (FDR) was computed using the q-value package (Storey and Tibshirani 2003) of Bioconductor. A q-value is a measure used to estimate the FDR, representing the proportion of false positives among significant results in multiple hypothesis-testing scenarios. Chromosome-wide and genome-wide Manhattan plots of P-values were generated to identify key regions associated with the pigmentation traits.

Association of most significant SNPs with other traits

The univariate analyses were conducted by including the genotype of the top significant SNP to analyse the associations with other traits. The ASReml-R function wald.asreml() was used to determine the significance of the genotype on the traits. Significance was defined as P < 0.05. The size of the SNP effect was also obtained from the analyses.

Descriptive statistics

The mean and coefficient of variance for each trait are presented in Table 4. NOSEP scores of individual animals ranged from one to five (Fig. 1a), with an overall mean score of 3.69. HVSP scores of individual animals ranged from 1.7 to 5.0, with an overall mean score of 4.12. NOSEP and HVSP were positively correlated (Fig. 1b).

Fig. 1.

Distribution of (a) nose-pigmentation score (NOSEP) and hoof-pigmentation score (HVSP), and (b) scatter plot of the NOSEP and HVSP scores.

Genome-wide association study

The genome-wide association analysis tested a total of 39 484 SNPs for their association with NOSEP and HVSP. For NOSEP, after adjusting for FDR, there was only one significant SNP (OAR19_33278780.1; q-value = 0.042, P = 0.000001) on Chromosome 19 (chr19:31614145, Oar_v4.0; Fig. 2, Table 5). OAR19_33278780.1 (rs419333175) SNP is located within the gene melanocyte-inducing transcription factor (MITF), which is known to be involved in pigmentation. No significant SNPs associated with HVSP were found (Fig. 3). However, OAR19_33278780.1 was the sixth-most significant SNP for HVSP, but remained below the significance threshold of the q-value of <0.05 (Table 5).

Fig. 2.

Genome-wide association study results for nose pigmentation in Corriedale sheep. (a) Histogram of P-values, (b) QQ-plot of P-values and (c) Manhattan plot of −log₁₀P-values. The P-values in the Manhattan plot are plotted by chromosome and the horizontal line indicates genome-wide significance (P = 5 × 10⁻⁶). Only one SNP was significant for NOSEP on Chromosome 19 (q-value = 0.042).

Fig. 3.

Genome-wide association study results for hoof pigmentation in Corriedale sheep. (a) Histogram of P-values, (b) qq-plot of P-values and (c) Manhattan plot of −log₁₀P-values. The P-values in the Manhattan plot are plotted by chromosome and the horizontal line indicates genome-wide significance (P = 5 × 10⁻⁶).

Of the lambs with NOSEP scores, there were 333 lambs with genotype GG, 32 with the genotype AG and only one lamb with the genotype AA of the OAR19_33278780 SNP. Lambs with the genotype GG at OAR19_33278780.1 SNP had a NOSEP mean score (3.75 ± 0.04 [s.e.]) 25% higher than for lambs with the genotype AG (3 ± 0.04), indicating that they had greater pigmentation (Fig. 4). Approximately 40% of lambs with the genotype AG had a NOSEP score of 3, while approximately half of the sheep with the genotype GG had a NOSEP score of 4 (Fig. 4b).

Fig. 4.

(a) Bar chart of mean and standard error for nose pigmentation (NOSEP) score grouped by genotypes (AA, n = 1; AG, n = 32; and GG, n = 333) at OAR19_33278780.1 SNP. (b) Density plot for NOSEP for AG and GG genotypes.

Effects of OAR19_33278780.1 genotypes on other measured traits

The genotype of OAR19_33278780.1 was added as a fixed factor in the univariate models and Wald tests were performed to determine effects of the genotype of this SNP on the wool, meat traits and visual scores. The genotypes for OAR19_33278780.1 were AA (n = 1), AG (n = 60) or GG (n = 736). The genotype at OAR19_33278780.1 was not significant for any of the visual scores. The genotype was significant for WWT (P = 0.01) and PWWT (P = 0.01), where GG was associated with higher weights. The genotype at OAR19_33278780.1 did not have a significant effect on the remaining traits.

An additional analysis was conducted by excluding the single lamb with the genotype AA. The genotype at OAR19_33278780.1 was significant for visual score, BACK (P = 0.02), and marginally significant for COL (P = 0.053, Table 6). Lambs with the genotype GG had 24.9% higher nose skin-pigmentation scores (darker pigmentation), 17.3% higher hoof-pigmentation scores (darker pigmentation), 7.0% lower scores for backline (more dipped) and 8.1% higher wool-colour scores (more yellow) than those with the genotype AG (Fig. 5, Table 6). The genotype at OAR19_33278780.1 was significant for WWT (P = 0.03) and close to significance level for PWWT (P = 0.06) and HSCW (P = 0.07) (Fig. 6, Table 6). Lambs with the genotype GG had 4.9% higher weaning weight, 3.9% higher post-weaning weight and 3.7% higher carcass weight (Fig. 3c, d, Table 6). The estimated means for these traits are given in Table 6.

Fig. 5.

Estimated means and standard errors for (a) nose-pigmentation (NOSEP) score, (b) hoof-pigmentation (HVSP) score, (c) back-conformation (BACK) score and (d) wool-colour (COL) score (yellowness) for sheep with genotypes AG and GG at OAR19_33278780.1.

Fig. 6.

Estimated means and standard errors for (a) weaning weight (WWT, kg), (b) post-weaning weight (PWWT, kg) and (c) hot standard carcass weight (HSCW, kg) for sheep with genotypes AG and GG at SNP OAR19_33278780.1.

Heritability estimates

The heritability estimates for different traits included in this study are presented in Table 7. NOSEP (h² = 0.58) and HVSP (h² = 0.51) were highly heritable, but the other visual scores were low to moderately heritable (h² = 0.17–0.44). Wool measures, GFW (h² = 0.80), FD (h² = 0.82) and CF (h² = 0.60) had high heritability estimates (Table 7), while staple length had a low heritability (h² = 0.20). The meat-production and eating-quality traits, PWWT, WWT, Ht, HSCW, EMW, PFat and GRFat, had high heritability estimates, whereas PEMD, EMD, EMA and IMF had moderate heritability estimates and, SF5 had a low heritability estimate.

Genetic and phenotypic correlations

The genetic and phenotypic correlations of NOSEP and HVSP with wool and meat traits, and visual scores, are presented in Table 7. NOSEP had genetic correlations with three traits (HVSP, PFat and FACE) and HVSP had genetic correlations with 13 traits (NOSEP, PWWT, WWT, Ht, HSCW, PEMD, EMD, EMA, PFAT, GRFat, IMF, CHAR and FACE).

NOSEP was highly positively correlated (genetically and phenotypically) with HVSP. NOSEP was highly positively correlated (genetically) with FACE score, but the phenotypic correlation was weak. HVSP was highly positively correlated (genetically) with FACE and highly negatively correlated (genetically) with CHAR. The phenotypic correlations between HVSP and visual scores, except for NOSEP, were close to zero.

NOSEP was moderately negatively correlated (genetically) with PFat. HVSP was highly negatively correlated (genetically) with Ht and EMD, and moderately negatively correlated (genetically) with PWWT, WWT, HSCW, PEMD, EMA, PFat, GRFat and IMF. The phenotypic correlations between NOSEP or HVSP and meat-production and meat-quality traits were either weak or close to zero.

SNP OAR19_33278780.1 (rs419333175) is associated with nose pigmentation

The GWAS pinpointed a single SNP (OAR19_33278780.1; chr19:31614145, Oar_v4.0) on Chromosome 19 associated with nose-pigmentation score (Fig. 2). Raadsma et al. (2013) reported a QTL at 20 cM (confidence interval = 0–53 cM) on Chromosome 19 linked to nose and hoof pigmentation, among other pigmentation traits, of Merino × Awassi F1 sheep. The authors proposed that MITF is located within this QTL on Chromosome 19. As OAR19_33278780.1 is located within an intron of the pigmentation gene, MITF, it is likely that the SNP also resides within the QTL. The current study provides further support for the involvement of the region on Chromosome 19 in pigmentation.

The GG genotype at OAR19_33278780.1 was associated with a 25% increase in nose-pigmentation scores compared with AG genotype. As the genotype at OAR19_33278780.1 does not follow Mendelian inheritance, the trait is likely to be affected by additional genetic variants. Other sources of variation could result from incomplete linkage disequilibrium between the SNP and the causative variant, as well as environmental factors and scoring inaccuracies.

Interestingly, only one individual possessed the genotype AA, highlighting the selective pressure favouring the G allele. This SNP was identified to be under positive selection in a study involving 74 diverse breeds of sheep (Kijas et al. 2012).

OAR19_33278780.1 may also be important for hoof pigmentation. OAR19_33278780.1 ranked among the top-10 most significant SNPs for HVSP, although it did not exhibit a significant q-value. Nevertheless, the genetic correlation between nose and hoof pigmentation was strong (Table 7). Previous studies on sheep have also identified OAR19_33278780.1 as a significant marker for pigmentation traits. This SNP was significantly associated with coat colour (white, black, ‘grey’ caused by white and black fibres, and brown) in a GWAS study of Finnsheep (Li et al. 2014). Additionally, QTLs for eye skin, ear skin, face skin, leg fibre, horn fibre and ear fibre pigmentation also mapped to Chromosome 19 in Merino × Awassi sheep (Raadsma et al. 2013). Taken together, MITF may be involved in hoof pigmentation, but the current study may have lacked sufficient observations or accurate scoring to detect significance for this trait.

Variants of MITF have been linked to white spotting in horses (Hauswirth et al. 2012; Negro et al. 2017; Henkel et al. 2019; Magdesian et al. 2020; Patterson Rosa et al. 2022) and water buffalo (Yusnizar et al. 2015). In horses, an intronic SNP (g.20147039C > T) was associated with white facial markings in the Spanish Purebreds, and white forelimbs in Menorca Purebreds (Negro et al. 2017). A large deletion including MITF was associated with white coat colour in a Holstein calf (Wiedemar and Drögemüller 2014). A de novo missense variant in Exon 8 of MITF was associated with white coat colour in a white American Standardbred foal (Dürig et al. 2017).

MITF, a transcription factor, plays a role in the migration and differentiation of neural crest cells, which serves as the pluripotent precursor of melanocytes (Widlund and Fisher 2003; Steingrímsson et al. 2004; Levy et al. 2006). Genetic variants of MITF influence pigmentation early in development by guiding the migration and differentiation of melanocyte precursors in specific regions. Consequently, skin and fibres lacking melanocytes appear white. In Corriedale sheep, the pigmented nose skin may be caused by a variant of MITF that specifies localised melanocyte development at the nose.

OAR19_33278780.1 significantly associated with back conformation, and weaning weight

The genotype at OAR19_33278780.1 was added to the univariate model to determine the effect on production traits. The genotype GG was also associated with 5.0% increased weaning weights and 7.3% reduced back-confirmation scores, indicating that back lines are slightly more dipped. Nevertheless, these effect sizes are small and may not have significant impacts on breeding programs.

The GG genotype was associated with 3.9% increase in post-weaning weights, 3.7% rise in carcass weights, and an 8.1% increase in wool colour, although these associations were not statistically significant. As the effect on weaning weight was significant, lambs with the GG genotype may exhibit higher early growth rates. However, the significance of the SNP diminished over time in subsequent weight measurements (PWWT and HSCW). Additionally, there might be insufficient observations to establish significance of the SNP effect for weight traits as the lamb ages and is exposed to more environmental influences. While lambs with the GG genotype tended to have more yellow wool, the effect size remained small.

MC1R may also be involved in pigmentation traits

A SNP on Chromosome 14 (Chr14:14283245) near melanocortin 1 receptor (MC1R) (Chr14:14,231,721–14,232,541; Oar_v3.1) was among the top 10 most significant SNPs for NOSEP and HVSP, but was not significant when the P-value was adjusted for FDR. MC1R is a 7-transmembrane G protein-coupled receptor expressed by melanocytes (Cieslak et al. 2011; Koseniuk et al. 2018) that acts as a pigmentation switch and stimulates eumelanin (black pigment) production when it binds to α melanocyte-stimulating hormone (Cieslak et al. 2011; Koseniuk et al. 2018). Although, the SNP near MC1R is not significant, it may still be involved with these pigmentation traits. This hypothesis could be investigated with more observations and objective measurement methods such as by using image analysis.

Pigmentation traits have some genetic correlations with meat traits

Darker hooves had moderate and high genetic correlations with meat and fat-quantity traits, which indicated reduced growth and fat coverage (Table 5). Some of these genetically correlated traits also had phenotypic correlations with hoof pigmentation; however, these were weak. Darker hooves were also significantly correlated with lower intramuscular fat percentage, indicating lower meat quality; however, the phenotypic correlation was weak. Nose pigmentation was not genetically correlated with wool- and meat-production traits, except for scanned fat depth. Darker noses had moderate genetic correlation with scanned fat depth. As the phenotypic correlations for NOSEP and HVSP with meat-production and meat-quality traits were weak or close to zero, any negative genetic associations that NOSEP and HVSP have on traits may be compensated by other genetic and environmental factors.

Corriedale sheep have mean trait measures similar to those of other breeds

These Corriedale sheep had good growth rates, with weaning weights (~5 months old) of 32.6 kg and post-weaning weights (~6–7 months) of 39.6 kg. This is greater than the WWT of Merino × lambs (27.8 kg; Mortimer et al. 2010) and Merinos (6.51–27.05 kg) and mean PWWT of Merinos (21.98–38.28 kg; Vaez et al. 1996; Huisman et al. 2008; Mortimer et al. 2017a, 2018). These Corriedale sheep had lower PWWT than those of Polled Dorset (45.2 kg; Gilmour et al. 1994).

The lambs were slaughtered at lighter weights (20.64 kg) than in similar trials with other meat breeds (21.6–23.1 kg; Mortimer et al. 2010, 2014) and Merino lambs (21.1 kg; Mortimer et al. 2017b, 2018). The EMA of Corriedale lambs was within the range of means reported in Merinos (11.6–13.1 mm; Mortimer et al. 2010, 2017b, 2018). The GRFat depth in Corriedales (7.46 mm) was lower than GRFat of Merinos (10.7 mm; Mortimer et al. 2018) and Merino × lambs (21.6 mm; Mortimer et al. 2010). IMF in Corriedales was within the range reported for Merinos and Merino × lambs (4.23–5.7%; Mortimer et al. 2010, 2014, 2018). The lower HSCW in these Corriedale may be attributed to the lower fat coverage.

The FD of Corriedale lambs (23 μm) was lower than an earlier reported mean of 25 μm in this breed (Brash et al. 1994), but greater than FD in yearling Merino sheep (16.9–18.1 μm; Huisman et al. 2008; Mortimer et al. 2017a). Mean GFW was 1.81 kg at approximately 5–6 months.

The Corriedale lambs had COL, SSTRC, CHAR and FACE scores similar to those of Merinos (COL: 2.37–2.85; SSTRC: 2.5–2.65; CHAR: 2.45–2.69; and FACE: 1.85–2.73; Hatcher and Preston 2018). The mean BDWR score of Corriedale lambs was lower than yearling, adult, and hogget Merino scores (2.09–2.64), indicating that Corriedale lambs are less wrinkled.