Genome-Wide Association Study for the Capacity to Skip the Dry Period in Dairy Goats

Abstract

Lactation is a challenging life stage for dairy animals, as they need to cope with milk production and, in most cases, simultaneous pregnancy. The dry period between two consecutive lactations can be a producer choice, based on, for instance, animal performance or a physiological requirement when animals dry off spontaneously. The goals of this research were to estimate genetic parameters and perform a genome-wide association study in Saanen goats for the capacity to skip the dry period between lactations to identify genes and QTLs underlying this trait. A total of 249 Saanen dairy goats had the length of their dry period determined over lactations, with some (n = 54) showing the capacity to skip the dry period, i.e., having a dry period of a single day. The estimated heritability for the capacity to skip the dry period was moderate (0.25, SE = 0.13). Three SNPs significantly associated with the capacity to skip dry period were identified, which are located close to the OSMR gene, reported to be associated with mammary involution, and a known QTL for cannon bone circumference. The three SNPs were also confined to a very conserved region on chromosome 20, which harbors several genes associated with milk-related traits. The OSMR gene seems to be a good candidate gene for the capacity to skip the dry period, and the genomic region where it is located appears to also be important for milk production traits.

1. Introduction

Milk production is a challenging field of study as it involves knowledge in several science fields, like biochemistry, cell biology, physiology, and morphology, among others. Milk is a secretion of carbohydrates, lipids, proteins, and fat in a watery medium. The mammary gland is a compound alveolar gland, where alveoli groups are arranged in lobules, and the internal epithelial layer of cells is responsible for milk production, which liberates milk components in the apocrine model for lipids and the merocrine model for protein and sugar. The mammary gland is capable of repeated cycles of growth, functional differentiation, and regression. The transition from lactating to regression status is mainly triggered by the secession of milk removal caused by suckling or milking. After this, the mammary gland will go through the apoptosis of milk production cells and tissue remodeling, and if there is no simultaneous pregnancy, the gland will return to a mature virgin structure [1]. These recurrences generate alternations between periods of milk production and dry periods.

Milk production is influenced by either the number or activity of epithelial cells [2]. Knight and Peaker [3] demonstrated that in dairy goats, the first part of the increase in milk yield during early lactation can be attributed to a rise in the number of mammary cells, followed by an enhancement in the productivity of individual cells. Conversely, the subsequent decline in milk production in late lactation is primarily due to a loss of mammary cells, with a later reduction in the production efficiency of the remaining cells.

There is no consensus about the need for a dry period, but apparently, its absence is less dramatic in goats than in cows, at least for production in the subsequent lactation. This can be explained by the goat’s great capacity to continue cell renewal during early lactation, which is equivalent to that observed during the dry period in cows. This goat capacity is attributable to two factors, significant mammary growth and somatotropin release stimulated by milking [2].

Some studies have shown the benefits of skipping dry periods, like those presented by Zobel et al. [4] who studied ketonemia in does. They showed that healthy animals presented shorter dry periods, and ketonemia was less present in a group of animals milked continuously.

A study comparing does that continuously milked to ones with a dry period showed that does that were milked continuously showed an improved postpartum energy balance in the first two weeks after kidding, high milk protein levels, no differences in the birth weight of kids, a reduction in the acute phase response (APR) during late gestation and early lactation, a relieved inflammatory condition close to parturition, and no difference in milk yield. However, the Somatic Cell Count was higher in does whose dry-off period was skipped, and they also had lower milk lactose levels at the beginning of lactation [5].

Caja et al. [6] reported negative consequences of skipping dry periods in goat colostrum quality (lower IgG concentrations) and milk yield. However, they also found no difference in mammary cell turnover among different lengths of dry-off periods, pointing to the possibility of using shorter dry periods.

Judging the value of dry periods in management systems involves balancing the amount of milk not produced during the dry period and the increased production in the following lactation [2]. The typical dry period for ruminants is approximately two months [5], although it varies by species. For dairy goats in Ontario, the mean dry period is around 40 days (SD ± 22) [4]. Reducing the dry period without compromising animal health and wellness could result in increased milk production over the animal’s productive lifespan. For instance, Caja et al. [6] reported that dairy goats dried off for 27 days were as productive (milk yield L/day) as those dried off for 56 days in the subsequent lactation, representing an additional 29 extra days of milk production. In a hypothetical scenario where an animal has six lactations during its productive life, with five possible dry-off periods, saving 29 days per dry-off period could result in an additional 145 lactating days throughout its productive life.

The goal of this study was to investigate the underlying genetic mechanisms associated with dry periods using a group of Saanen goats classified as having no dry periods, from a farm that implements extended lactations in its management system.

2. Materials and Methods

A total of 249 Saanen dairy goats from a farm located in Ontario, Canada (44°27′ N, 78°32′ W), had the length of their dry period determined over lactations, with some (n = 54) showing the capacity to skip the dry period, i.e., having a dry period of a single day. This farm is quite unique in the sense that, for many years, the farmer has implemented a protocol to extend the doe lactations for as long as possible, meaning that does are targeted for milking until parturition, in addition to aiming for no dry period between lactations if the does will not stop or immediately resume milk production after parturition. Under this management, after kidding, some does resume milk production just after 1 day, whereas other does require a longer dry period before lactation is restarted. This variation shows an inherent capacity in certain does to skip the dry period.

The climate parameters for the location of this farm—including relative humidity (%) and temperature (°C) at 2 m above ground, recorded from 2004 to 2015—were obtained from the NASA POWER | DAV platform [7], and monthly average values were calculated and are shown in Supplementary Figure S1.

The number of days dry (dry period) for each doe was calculated as the difference between the potential days in milk (PDIM) and the real DIM recorded for a specific lactation, as shown in Figure 1. The PDIM was the difference in days between two consecutive parturitions (e.g., the date of parturition n + 1 minus the date of parturition n was the PDIM for lactation n). The calculation of the dry period was performed for all does’ lactations, but only the lowest dry period was kept to define which does had the capacity to skip the dry period. To improve the power of detecting significant associations, 15 does with intermediary dry period phenotypes were removed from the dataset, i.e., does that had dry periods higher than 1 but lower than 16. Finally, the dry period values were categorized into a binary trait, where 1 represents dry period = 1 day (no dry period; i.e., the doe showed the capacity to skip the dry period), and 0 represents dry periods ≥16 days (normal dry periods).

The average age for the animals at the moment of parturition, before the lactation with the smallest number of days dry, was 2.2 years (SD ± 1.28). A summary of milk production traits (lactation count, total DIM, total milk, fat, and protein yields) for the does analyzed is provided in Supplementary Table S1.

These animals were genotyped using the GoatSNP50 Bead Chip from Illumina Inc. San Diego, CA, USA, and the genotypic data underwent quality control using the PLINK 1.9 software [8,9]. Markers with a minor allele frequency (MAF) below 0.01, those with missing call rates above 0.1, and variants failing Hardy Weinberg Equilibrium at a p-value threshold of 1 × 10⁻⁵ were excluded. Samples with missing call rates exceeding 0.1 were also removed. After applying these filters, 44,649 SNPs were retained.

To ensure data integrity, the dataset was checked for duplicated samples using PLINK 2.0 [9,10], applying a kinship coefficient cutoff of 0.354 [11,12]. The SNP positions were mapped to the ARS1 (version 1) goat reference genome assembly [13]. Additionally, the population structure was assessed through a principal component analysis (PCA), which was also conducted in PLINK 1.9 [8,9].

The ASReml 4.2ni software [14,15] was used to test the significance of possible systematic environmental effects included in the model for the estimation of genetic parameters and for the genome-wide association study. The effects of the birth year of the doe (classification variable) and days in milk just before parturition (linear covariate) were significant (p < 0.05) and were kept in the final model to estimate the variance components and heritability of the trait using ASReml 4.2ni:

$$\mathbf{y}=\mathbf{W}\mathsf{\gamma }+\mathbf{Z}\mathbf{a}+\mathbf{e}$$

(1)

where $\mathbf{y}$ is the vector of trait phenotypes (0 or 1); $\mathsf{\gamma }$ is the vector for fixed effects (birth year and days in milk), and $\mathbf{W}$ is its incidence matrix; $\mathbf{a}$ is the vector of random additive polygenic genetic effects and $\mathbf{Z}$ is its incidence matrix and $\mathbf{e}$ is the vector of the random residual effects. Both a and e were assumed to follow normal distributions with mean zero and covariance matrices equal to ${\mathbf{A}\mathsf{\sigma }}_{\mathrm{a}}^2$ and I${\mathsf{\sigma }}_{\mathrm{e}}^2$, respectively, where A is the additive numerator relationship matrix.

The heritability of the trait was estimated as

$${\mathrm{h}}^2={\displaystyle \frac{{\mathsf{\sigma }}_{\mathrm{a}}^2}{{\mathsf{\sigma }}_{\mathrm{a}}^2{+\mathsf{\sigma }}_{\mathrm{e}}^2}}$$

(2)

For the genome-wide association study, a mixed linear model association analysis with the leave-one-chromosome-out (LOCO) approach, implemented using the GCTA v1.94.1 software package [16], was used. The statistical model applied was as follows:

$$\mathbf{y}=\mathbf{W}\mathsf{\gamma }+\mathbf{X}\mathsf{\beta }+\mathbf{Z}\mathbf{u}+\mathbf{e}$$

(3)

where y represents the vector of phenotypic observations, with 0 indicating a normal dry period and 1 indicating animals that skipped the dry period. $\mathbf{W}$ is the incidence matrix for the fixed effect, $\mathsf{\gamma }$ (birth year and days in milk). The vector $\mathbf{X}$ corresponds to the recoded genotypes (0, 1, or 2) of the single-nucleotide polymorphism (SNP) being tested, while $\mathsf{\beta }$ represents the SNP allele substitution effect. The matrix $\mathbf{Z}$ is the incidence matrix for the random additive polygenic effects (u), and $\mathbf{e}$ is the vector of random residual effects. Both $\mathbf{u}$ and $\mathbf{e}$ were assumed to follow normal distributions with mean zero and covariance matrices equal to ${\mathbf{G}\mathsf{\sigma }}_{\mathrm{u}}^2$ and I${\mathsf{\sigma }}_{\mathrm{e}}^2$, respectively, where $\mathbf{G}$ is the genomic relationship matrix (GRM) derived from SNP markers [17], excluding the chromosome where the SNP being tested is located (LOCO approach).

SNP associations were declared significant when their p-values were smaller than a Bonferroni correction threshold considering the expected number of independent chromosome segments (Me). This threshold was obtained by dividing the significance level ($\alpha =0.05$) by Me [18]:

$$\mathrm{M}\mathrm{e}={\displaystyle \frac{2\times \mathrm{N}\mathrm{e}\times \mathrm{L}}{\log 10\left(\mathrm{N}\mathrm{e}\times \mathrm{L}\right)}}$$

(4)

where $\mathrm{N}\mathrm{e}$ is the effective population size (assumed to be 113 [19]), and $\mathrm{L}$ is the length of the autosomal genome in Morgan calculated from the ARS1 assembly (version 1), summing up the length of all autosomes (from 1 to 29).

Quantile–quantile (Q-Q) plots were used to compare the observed distribution of -log (p-value) to the expected distribution under the null hypothesis of no association for the trait under study. The inflation factor, lambda ($\lambda$), was also calculated to assess potential population stratification. A Manhattan Plot was built using R packages dplyr v1.1.4 [20], ggplot2 v3.5.1 [21], and ggrepel v0.9.5 [22] according to [23]. The SNPs were mapped to the ARS1 goat reference genome [13].

The SNPs significantly associated with the phenotype were used in the R package GALLO v1.4 [24] to perform a search for genes and QTLs in an interval of 50,000 bases upstream and downstream, which was based on the average distance and linkage disequilibrium between adjacent SNPs for the Saanen breed [19]. The genes found within this range were used to perform a functional enrichment analysis in the online version of the g:Profiler v.e112_eg59_p19_25aa4782 software [25].

The variance explained by the significant SNPs was also estimated as follows [26]:

$$\mathrm{V}\mathrm{a}=2\mathrm{p}\mathrm{q}{\mathsf{\beta }}^2$$

(5)

where $\mathrm{p}$ is the frequency of the reference allele, $\mathrm{q}$ is the frequency of the alternative allele, and $\mathsf{\beta }$ is the estimated allele substitution effect.

3. Results

3.1. Variance Components and Heritability Estimation

The estimated additive genetic variance (${\sigma }_a^2$) and heritability of the capacity to skip the dry period were 0.04 (SE = 0.02) and 0.25 (SE = 0.13), respectively.

3.2. GWAS

The qq-plot in Figure 2 shows no inflation of the test statistic, with a genomic inflation factor (λ) of 0.96, indicating that the association test results are well calibrated and free from significant stratification effects. Additionally, the PCA (Figure 3) did not show any population structure within the sample, confirming the absence of stratification that could bias the genomic analyses.

Three significant SNPs on chromosome 20 associated with the capacity to skip the dry period were identified in this population of Saanen goats. Figure 4 shows a Manhattan Plot with these significant variants, while

Table 1. Significant SNPs after Bonferroni correction considering the number of genome-wide independent chromosome segments ($\alpha =0.05$).

Chr	SNP rsID	Position	A1	A2	MAF	$$\mathsf{\beta }$$	Va	%Va
20	rs268262366	32099030	G	A	0.060	0.357	0.014	0.352
20	rs268275254	34904494	A	G	0.183	0.232	0.016	0.391
20	rs268283969	35561292	A	G	0.068	0.344	0.015	0.366

A1: minor allele; A2: major allele; MAF: minor allele frequency; $\mathsf{\beta }$: allele substitution effect; Va: variance explained by the SNP; %Va: percentage of additive genetic variance (${\sigma }_a^2$) explained by each SNP individually (%Va = Va/${\sigma }_a^2$).

details their rsID, positions, allele frequencies, and variance explained by each one individually.

3.3. Gene and QTL Search

A search within a range of 50k BP downstream and upstream of the significant SNPs found two genes—the OSMR gene (ENSCHIG00000009367) near the SNP rs268283969 and the gene for long intergenic non-coding RNA (lincRNA), ENSCHIG00000003885—near SNP rs268275254. QTL 223248, for cannon bone circumference, was also identified in proximity to SNP rs268262366.

All three significant SNPs are located within the same region of chromosome 20. The first SNP (rs268262366) is located at position 32099030 bp, while the last SNP (rs268283969) is located at 35561292 bp, spanning a total range of 3462262 bp. Within this region, extended to an additional 60 kb, 23 different genes (

Table 2. Genes found in the extended window.

Gene ID	Gene Name	Gene_Biotype
ENSCHIG00000000538	RIMOC1	protein coding
ENSCHIG00000000675	SNORD72	snoRNA
ENSCHIG00000001827	U2	snRNA
ENSCHIG00000002172	-	protein coding
ENSCHIG00000003882	-	lincRNA
ENSCHIG00000003885	-	lincRNA
ENSCHIG00000009367	OSMR	protein coding
ENSCHIG00000009529	DAB2	protein coding
ENSCHIG00000012124	OXCT1	protein coding
ENSCHIG00000012521	-	protein coding
ENSCHIG00000014444	RPL37	protein coding
ENSCHIG00000014791	C9	protein coding
ENSCHIG00000016536	C7	protein coding
ENSCHIG00000016815	PLCXD3	protein coding
ENSCHIG00000017495	PRKAA1	protein coding
ENSCHIG00000018413	FYB1	protein coding
ENSCHIG00000018928	GHR	protein coding
ENSCHIG00000020642	TTC33	protein coding
ENSCHIG00000020994	CARD6	protein coding
ENSCHIG00000021210	C6	protein coding
ENSCHIG00000021750	FBXO4	protein coding
ENSCHIG00000022858	RICTOR	protein coding
ENSCHIG00000023835	PTGER4	protein coding

) were identified, which hereafter will be called the extended window. This window was extended to 60 kb instead of 50 kb based on the fact that the GHR gene, which is located within 60 kb of the significant SNPs, is known to be associated with cannon bone circumference and milk production, and there is still considerable expected linkage disequilibrium at this extended distance.

Notably, this region on chromosome 20 in goats is highly conserved and shares significant homology with the corresponding region in cattle. It occupies the same chromosome position and contains nearly identical genes, as illustrated in Figure 5.

3.4. Functional Enrichment Analysis

The functional enrichment analysis of the OSMR and ENSCHIG00000003885 genes and the 23 genes in the extended window returned the enriched terms in

Table 3. Significantly enriched terms associated with the OSMR gene and long intergenic non-coding RNA (lincRNA) ENSCHIG00000003885.

Source	Term Name	Term ID	FDR *
GO:MF	oncostatin-M receptor activity	GO:0004924	0.001732702
GO:MF	cytokine receptor activity	GO:0004896	0.013572831
GO:MF	growth factor binding	GO:0019838	0.013572831
GO:MF	immune receptor activity	GO:0140375	0.013572831
GO:BP	oncostatin-M-mediated signaling pathway	GO:0038165	0.005157358
GO:CC	oncostatin-M receptor complex	GO:0005900	0.001577731
GO:CC	apical plasma membrane	GO:0016324	0.04212543
GO:CC	receptor complex	GO:0043235	0.04212543
GO:CC	apical part of cell	GO:0045177	0.04212543
GO:CC	plasma membrane signaling receptor complex	GO:0098802	0.04212543
HP	cutaneous amyloidosis	HP:0012309	0.029205386
HP	lichenification	HP:0100725	0.029205386

GO: Gene Ontology; MF: Molecular Function; BP: Biological Process; CC: Cellular Component. * Benjamini and Hochberg false discovery rate threshold.

and

Table 4. Significantly enriched terms for the genes in the extended window.

Source	Term Name	Term ID	FDR *
GO:MF	cytokine receptor activity	GO:0004896	0.0374
GO:MF	histone H2B kinase activity	GO:0140998	0.0374
GO:MF	histone H2BS36 kinase activity	GO:0140823	0.0374
GO:MF	immune receptor activity	GO:0140375	0.0374
GO:MF	[acetyl-CoA carboxylase] kinase activity	GO:0050405	0.0374
GO:MF	[hydroxymethylglutaryl-CoA reductase (NADPH)] kinase activity	GO:0047322	0.0374
GO:MF	succinyl-CoA:3-oxo-acid CoA-transferase activity	GO:0008260	0.0374
GO:MF	oncostatin-M receptor activity	GO:0004924	0.0374
GO:MF	CoA-transferase activity	GO:0008410	0.0493
GO:MF	prostaglandin E receptor activity	GO:0004957	0.0493
GO:MF	AMP-activated protein kinase activity	GO:0004679	0.0493
GO:CC	membrane attack complex	GO:0005579	0.0000
GO:CC	pore complex	GO:0046930	0.0001
GO:CC	plasma membrane protein complex	GO:0098797	0.0111
GO:CC	growth hormone receptor complex	GO:0070195	0.0194
GO:CC	oncostatin-M receptor complex	GO:0005900	0.0310
KEGG	systemic lupus erythematosus	KEGG:05322	0.0097
KEGG	coronavirus disease—COVID-19	KEGG:05171	0.0097
KEGG	complement and coagulation cascades	KEGG:04610	0.0097
HP	decreased circulating terminal complement component concentration	HP:0033057	0.0000
HP	abnormality of complement system	HP:0005339	0.0022
HP	complement deficiency	HP:0004431	0.0022
HP	recurrent meningococcal disease	HP:0005381	0.0022
HP	recurrent Neisserial infections	HP:0005430	0.0022
HP	recurrent Gram-negative bacterial infections	HP:0005420	0.0375
HP	decreased circulating complement C9 concentration	HP:0012308	0.0387
HP	decreased circulating complement C6 concentration	HP:0033059	0.0387
HP	decreased circulating complement C7 concentration	HP:0033058	0.0387

GO: Gene Ontology; MF: Molecular Function; CC: Cellular Component; KEGG: Kyoto Encyclopedia of Genes and Genomes; HP: Human Phenotype Ontology. * Benjamini and Hochberg false discovery rate threshold.

, respectively.

4. Discussion

An estimated moderate heritability for the capacity to skip the dry period, along with substantial additive genetic variance, indicates the possibility of genetically selecting for this trait in the herd evaluated. However, data from other herds would be needed to validate these results and make broader and stronger inferences.

The search for more efficient milk production may also focus on practices that increase lactation persistency and shorten the dry period [2]. To reach the dry stage, animals need to undergo mammary gland involution, which depends on the apoptosis process [1,28]. The involution of a mammary gland is an intricate process that demands the extensive apoptosis of the secretory epithelium and the arrival of immune cells like macrophages and others to eliminate milk residuals and dead cell debris [29]. This complex process of involution is coordinated by signaling molecules, proteins, and hormones [28].

Based on a biochemical and histological analysis of mice, Lund et al. [30] divided the post-lactation involution of the mammary gland into two different stages: the first, still with milk production and the beginning of epithelium cell apoptosis, and the second, when severe tissue remodeling and proteinase synthesis occur. Lund et al. [30] showed that the lobulo-alveolar structures were unimpaired for up to three days after weaning. However, days 4, 5, 7, and 10 are marked by alveoli disintegration of 60%, 70%, 85%, and more than 95%, respectively. The period between days two and three is marked by an increased number of apoptotic epithelial cells in the alveolar lumina. However, until this period, no visible alterations were noted between mesenchymal and epithelial compartments. Conversely, the second involution stage is marked by severe tissue restructuring starting on day four that disintegrates the lobulo-alveolar structure.

The Oncostatin M receptor gene (OSMR) is the closest gene to one of the significant SNPs. Its protein product is located on the external side of the plasma membrane [31], and it functions as a receptor for the cytokine Oncostatin M (OSM), a member of the Interleukin-6 family [32].

The peak signaling of OSMR/OSM in mice mammary glands takes place on about the third day after weaning [33], which coincides with the end of the first stage of mammary involution and the beginning of the second stage [30,33]; thus, on about days 3 and 4, the main functions of OSM are heightened, including the promotion of apoptosis in the mammary epithelial cell; the downregulation of milk protein genes like β-casein and whey acid protein (WAP); and the metalloproteinase-dependent collapse of basement membranes, which are involved in the reconstruction of mammary gland tissue compartments [33].

The first stage of mammary gland involution, which lasts for up to two days after weaning in mice, is considered “reversible” if suckling is resumed [28]. After this period, if lactation is not recommenced, the animal goes through the second stage of mammary involution, which is not reversible.

OSM/OSMR genes have been mostly studied in mice; however, these genes have been found to be upregulated in the mammary glands of goats in the dry period when compared to the lactation period [34], showing that they likely also play some role in the mammary involution process in goats.

It has also been shown that after 10 days of the weaning, even OSMR-deficient mice have an increase in apoptosis, which can be a compensation caused by signal transducer and activator of transcription 3 gene (STAT3) independent proapoptotic signals [33]. It has also been reported that the highest OSMR activity is STAT3-dependent [33]. STAT3 is a member of the STAT family. Members of this family are activated by phosphorylation induced by cytokines and growth factors, such as leukemia inhibitor factor (LIF), and then, they can activate the transcription of several genes in response to cell stimuli, therefore participating in cellular processes like cell growth and apoptosis [35]. In goats, the STAT3 gene is located on chromosome 19 [36]. In summary, LIF can promote the activation of the STAT3 gene [37], which, in turn, can stimulate OSM transcription during the first stage of involution [28,33], and OSM and OSMR are the main proteins responsible for STAT3 activation during the second phases of involution [33,38]. The production of LIF is very low during lactation, but after suckling stops, with milk stasis, its level grows rapidly [28]; however, its action is dependent on its receptor (LIFR), which is also present in the significant genomic region in the current study (Figure 5), highlighting the importance of this region for mammary gland involution. Another example of the interaction between genes present in this region is that the OSM gene can use the LIFR, in addition to OSMR, and when this happens, the effect is different when using OSMR [39].

OSM/OSMR absence does not affect apoptosis during the first 48 h (the first stage of mammary gland involution) but has a marked effect on the decline of apoptosis during days three and four (the second stage of mammary gland involution) [33].

Another effect of OSM/OSMR during mammary gland involution in mice comes through the downregulation of active STAT5 [33], which increases the expression levels of BIM (Bcl-2-interacting mediator of cell death), highlighted as one of the regulators of the mitochondrial apoptosis pathway in the second phase of involution [40].

The participation of the OSM/OSMR gene in other processes of tissue degeneration/remodeling has also been shown in cattle follicle atresia, luteolysis, and ovulation [41]. Other authors have also claimed that the OSMR gene is also involved in the regulation of apoptosis in the pituitary glands of goats [42].

Even though only the OSMR gene was found within a range of 50 k from one of the significant SNPs, the importance of this region—located on chromosome 20 of goats and cattle and on chromosome 16 in sheep for lactation-related traits—should be emphasized. This region has been reported to be associated with mastitis resistance in both sheep and cattle [43,44]. In sheep, 8 out of 14 candidate genes associated with mastitis resistance (C6, C7, C9, PTGER4, DAB2, CARD6, OSMR, and FYB) are located within this region [43]. Similarly, in cattle, the GHR, OXCT1, C6, C7, C9, CARD6, DAB 2, OSMR, PRLR, and C1QTNF3 genes have been linked to clinical mastitis and are mapped to this region [44] (Figure 5).

Genes found in this region (Figure 5) have also been reported to be associated with milk-related traits; for instance, the GHR gene was found to be related to protein, fat, casein, lactation persistency, somatic cell score, Rennet coagulation time, and curd firmness 30 min after rennet addition [45,46,47,48,49], and the FYB, RICTOR, and C9 genes are associated with protein yield [45]; the OSMR gene with milk yield [46]; and the CCDC152 gene with lactation persistency [49], which reinforces the importance of this portion of the genome for milk production.

The cannon bone circumference QTL was found to be co-localized with one of the significant SNPs. Cheng et al. [50] found a significant association between the cannon circumference phenotype and specific mutations in the OSMR and GHR genes, both located in the significant region on chromosome 20 found in the current study. The participation of OSM/OSMR in bone resorption and formation through OSM/LIFR has already been demonstrated [39] and is fundamental for the process of bone growth [51].

In a comparison of differentially expressed genes between lactation and dry periods in dairy goats, Guan et al. [34] also found an upregulation of genes related to immune response, among them, genes belonging to complement cascade like the C6 and C7 genes, which were also found in the interval between the first and last significant SNP in the current study. These authors found that the pathway complement and coagulation cascades were enriched, and these were also enriched in the current study (KEGG:04610). Additionally, other terms related to immune response, like immune receptor activity (GO:0140375), complement deficiency (HP:0004431), membrane attack complex (GO:0005579), and pore complex (GO:0046930), among others, were also enriched in the current study.

Contrary to the upregulation of genes related to apoptosis and immune response, several genes involved in the production of milk components, like proteins and lipids, and the transportation of carbohydrates, amino acids, and minerals are downregulated as lactation ceases [34]. Some terms related to lipid metabolism were found to be enriched in the current study, such as [acetyl-CoA carboxylase] kinase activity (GO:0050405), [hydroxymethylglutaryl-CoA reductase (NADPH)] kinase activity (GO:0047322), and succinyl-CoA:3-oxo-acid CoA-transferase activity (GO:0008260).

Altogether, these results highlight the importance of the OSMR gene and the region where it is located for the length of dry periods in goats, as well as for other milk-related traits, fertility, growth, and health traits.

Capuco and colleagues [2] state that due to the complexity of the regulation of apoptosis and cell proliferation, new methodologies aimed at reducing the dry period and extending lactation can emerge from novel approaches, which include selecting for favorable genetic polymorphisms. In this context, the OSMR gene can be considered an important candidate gene when aiming to alter the dry period length.

5. Conclusions

The duration of the dry period determines how much time the animal will be not producing milk, which impacts the overall sustainability. The possibility of skipping the dry period has been assessed in several studies, and there is no consensus about its use. Regardless whether the objective is to skip the dry period, make it shorter, or even lengthen it, understanding the genetic makeup influencing the variation in this trait is important. The capacity to skip the dry period seems to be moderately heritable and, therefore, possible to be genetically selected for. The OSMR gene seems to be a promising candidate gene related to the length of the dry period, and the region where it is located on chromosome 20 has been previously associated with milk production, growth, and health traits. Future research using bigger datasets is warranted to validate the findings in this study and develop tools for producers to select for a dry period length that better suits their production and management systems.