Genome-wide assessment of runs of homozygosity and inbreeding in Inner Mongolia cashmere goats reveals candidate genes for economic traits

Abstract

Introduction

To facilitate the conservation and sustainable utilization of the Inner Mongolia Cashmere Goat (IMCG)-an important genetic resource and economically valuable breed-this study performed a genome homozygosity assessment of its Erlangshan subtype population, analyzed the distribution characteristics of runs of homozygosity (ROH), and compared inbreeding coefficients estimated using different methods.

Methods

A total of 384 individuals were genotyped using the Illumina Goat SNP70K BeadChip. PLINK software was employed to conduct ROH analysis, estimate genomic inbreeding coefficients, and identify ROH islands based on the obtained SNP data.

Results

The results revealed 3,622 ROH segments in the population, with an average total ROH length of 71.48 Mb per individual. The inbreeding coefficients based on ROH (F_ROH) and observed homozygosity (F_HOM) were 0.0290 and −0.0126, respectively, showing a strong positive correlation (r = 0.97). In addition, 14 ROH islands were identified on chromosomes 2, 5, 7, 8, 9, 13, 17, 19, 24, 25, and 26. Within these regions, 238 genes were annotated, including key genes such as HDAC1, MIR21, DNASE1L2, and ACACA, which may be implicated in reproductive traits, hair follicle development, and milk fat metabolism.

Conclusion

This study systematically characterized ROH patterns in IMCG, indicating a generally low level of genomic inbreeding in the population, though certain individuals still require attention. The identification of genes related to economically important traits further reveals the genetic distinctiveness and breeding potential of IMCG, providing valuable insights for informing future conservation and breeding strategies.

1. Introduction

China is rich in indigenous genetic resources of cashmere goats. Among them, the Inner Mongolia Cashmere Goat (IMCG) represents a distinctive breed consisting of three subtypes: Albas, Erlangshan, and Alashan (1). Known for producing high-quality meat and fine cashmere, IMCG is a vital component of China’s cashmere goat genetic resources (2). To date, research has primarily focused on selective breeding for cashmere traits, differential gene expression, and molecular regulation of hair follicle development, while analyses of runs of homozygosity (ROH) patterns remain limited (1). ROH analysis based on genomic data can reveal the extent of inbreeding, recent population bottlenecks, and signatures of directional selection. Therefore, this study addresses this knowledge gap by systematically describing the occurrence and distribution of ROH in the IMCG population using medium-density single nucleotide polymorphisms (SNPs) chip data, thereby laying a foundation for subsequent genetic studies.

In genomic research, ROH refers to continuous homozygous DNA segments in diploid organisms that are inherited identically by descent from common ancestors (3). Because genetic recombination breaks such homozygous fragments over generations, ROH length serves as an important indicator for tracing the timing of inbreeding events: long ROH segments typically reflect recent inbreeding, whereas shorter ones indicate more ancient common ancestry (4). With the marked reduction in genotyping costs and the rapid increase in genomic data, goat breeding has entered the era of genomic selection. Consequently, methods for estimating inbreeding coefficients are gradually shifting from traditional pedigree-based approaches to genomic computations based on ROH (5). Compared with indirect pedigree-based estimates, the genomic inbreeding coefficient derived from ROH (F_ROH) offers clear advantages, especially when pedigree records are incomplete or missing.

ROH analysis also facilitates the identification of genomic regions with high frequencies of homozygosity, termed ROH islands. Detailed examination of these regions can provide deeper insights into genomic intervals under selective pressure, thereby revealing key information about population genetic adaptation and evolutionary history (6). The detection of ROH islands not only complements genome-wide association studies but also helps uncover species-specific important genes. In recent years, genome-wide analyses of ROH have been extensively applied across diverse livestock and poultry species to characterize population structure, estimate genomic inbreeding, and detect signatures of selection. For instance, Ma et al. (7) employed a genome-wide ROH approach in ovine populations, revealing breed-specific inbreeding patterns and detecting candidate genes associated with wool quality and reproductive efficiency. Hervás-Rivero et al. (8) through the mapping of ROH islands in autochthonous Spanish beef cattle breeds, uncovered genes related to muscle growth (e.g., MSTN) and fertility under shared ancestral selection pressures. Rostamzadeh Mahdabi et al. (9) via comparative ROH island analyses in indigenous and commercial chickens, precisely pinpointed candidate loci associated with disease resistance, heat tolerance, and reproduction. These cross-species applications demonstrate the powerful role of ROH analysis in elucidating the genetic basis of economically important traits and guiding breeding strategies. Therefore, this study aims to systematically identify and characterize ROH patterns in the IMCG (Erlangshan subtype) population using medium-density SNP chip data. We assess its inbreeding status, screen potential candidate genes within ROH islands associated with key production traits, and provide foundational genomic information to support future conservation and breeding programs.

2. Materials and methods

2.1. Experimental animals and its care

All animal experiments were performed by the Guidelines for Experimental Animals of the Ministry of Science and Technology (Beijing, China) and were approved by the Scientific Research and Academic Ethics Committee of Inner Mongolia Agricultural University and the Biomedical Research Ethics of Inner Mongolia Agricultural University (Approval No. [2020] 056).

2.2. Animal origin and DNA extraction

A total of 384 IMCG (Erlangshan subtype) individuals (104 rams and 280 ewes) were sampled from the Erlangshan Pasture (108°31′E, 41°26′N, altitude 1,248 m) of Inner Mongolia Beiping Textile Co., Ltd. After ear tissue collection from all individuals, genomic DNA was extracted using the standard phenol-chloroform method (10). DNA purity (260/280 nm ratio) was assessed with a Nanodrop 2000 spectrophotometer (Thermo Fisher Scientific Inc., United States). Qualified DNA samples were stored at −80 °C for subsequent genotyping.

2.3. Genotyped and quality control

Genotyping of qualified samples was performed using the Illumina Goat SNP70K BeadChip (Inner Mongolia Agricultural University, China). Genotype quality control was conducted with PLINK (V1.90) software (11). Single nucleotide polymorphisms (SNPs) were removed if they met any of the following criteria: (i) call rate < 90%; (ii) minor allele frequency < 0.05; (iii) Hardy–Weinberg equilibrium p-value < 1 × 10⁻⁶; or (iv) individual genotype missing rate >10%. Additionally, SNPs located on sex chromosomes and those with ambiguous chromosomal positions were also excluded. After quality control, 50,933 SNPs from 380 animals were retained for subsequent analyses. The SNPs dataset has been deposited in the Figshare database.¹

2.4. Runs of homozygosity detection and classification

PLINK (V1.9) software (11) employs a sliding window algorithm to analyze the genome and identify ROH. In consideration of the characteristics of the medium-density Illumina Goat SNP70K BeadChip, the following parameter set was applied in this study: (1) a sliding window of 50 SNPs across the genome; (2) to minimize the number of false positive ROH, the minimum number of SNPs constituting the ROH was calculated using the method proposed by Lencz et al. (12):

$$l=\frac{{log}_e\alpha /n_s\times n_i}{{log}_e(1-\mathit{het})}$$

where $\alpha$ is the percentage of false positive ROH (set to 0.05 in this study), $n_s$ is the number of SNPs per individual, $n_i$ is the number of individuals, $\mathit{het}$ is the heterozygosity across all SNPs; (3) up to one missing SNP and one heterozygous genotype were allowed per ROH to accommodate potential genotyping errors (13); (4) a maximum gap between consecutive SNPs of 1 Mb and minimum SNP density was set to 1 SNP per 100 kb; (5) to distinguish true autozygous segments from short homozygous tracts likely arising from linkage disequilibrium (LD), we opted to set a minimum length threshold of 1 Mb for ROHs instead of performing LD pruning (14). The parameter set chosen here represents a balanced strategy widely used in livestock genomics research for medium-density SNP array data.

2.5. Runs of homozygosity distribution and inbreeding coefficients

First, the mean number of ROH (MN_ROH), mean length of ROH (ML_ROH), and the total number of ROH was estimated for each animal. The number of ROH on different chromosomes and the percentage of chromosomes covered by ROH were also calculated. This coverage percentage for each chromosome was calculated by dividing the total length of ROH on that chromosome by the full length of the corresponding chromosome as per the ARS1.2 goat reference genome assembly. Next, the lengths of all ROH were categorized into four categories: 1–5 Mb, 5–10 Mb, 10–20 Mb, and >20 Mb, and the number of ROH in each category on each chromosome was counted. In addition, we estimated and analyzed two measures of inbreeding: the coefficient based on observed homozygosity (F_HOM) and that based on ROH (F_ROH). F_ROH was estimated for each individual using PLINK according to McQuillan et al. (15):

$$F_{\mathit{ROH}}=\frac{\sum L_{\mathit{ROH}}}{L_{\mathit{\text{auto}}}}$$

where $L_{\mathit{ROH}}$ is the total length of all ROH in an individual, and $L_{\mathit{\text{auto}}}$ refers to the autosomal genome length covered by SNPs included in the array. F_HOM was assessed based on the proportion of homozygotes. Pearson correlation coefficients and their corresponding significance test values were calculated using the cor.test function in the R (V4.4.1) (16). To calculate correlation coefficients, F_ROH coefficients were also estimated for four length categories of ROH (F_{ROH(1–5 Mb)}, F_{ROH(5–10 Mb)}, F_{ROH(10–20 Mb)}, and F_{ROH(>20 Mb)}).

2.6. Detection of ROH islands and gene annotation

To identify the genomic regions most frequently associated with ROH, we counted the number of times SNPs were detected in ROH across all individuals to calculate the percentage of SNPs present in ROH. In this study, the top 1% of SNPs observed in ROH was chosen as the threshold for identifying the genomic region most frequently associated with ROH. A series of neighboring SNPs above this threshold formed a genomic region, which we called an ROH island.

To annotate the genes within the ROH islands, we mapped their genomic coordinates to the goat reference genome (ARS1.2, GCF_001704415.2) using NCBI’s genomic data viewer.² Subsequently, all candidate genes located in these islands were subjected to functional enrichment analysis using the DAVID tool (V6.8) (17) to identify significant Gene Ontology (GO) terms and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways. Additionally, to further dissect the functional associations among candidate genes, a protein–protein interaction (PPI) network was constructed using the STRING database (v12.0) (18). Interactions with a combined score ≥ 0.4 (medium confidence) were retained, and hub genes were identified based on node degree centrality.

3. Results

3.1. Distribution of runs of homozygosity

A total of 3,622 ROH segments were identified from the 380 individuals. On average, each individual carried 9.53 ROHs (range: 1–37), and the mean cumulative ROH length per individual was 71.48 Mb. All individuals possessed at least one ROH segment ≥1 Mb in length. Significant differences were observed in ROH number and coverage rate among chromosomes: chromosome 1 harbored the highest number of ROH segments (229 segments), yet exhibited the lowest coverage rate (6.23%); in contrast, chromosome 26 showed the highest coverage rate (18.24%), while chromosome 27 contained the fewest ROH segments (65 segments) (Figure 1). The longest ROH was located on chromosome 29 (28.24 Mb), and the shortest on chromosome 1 (1.95 Mb).

Figure 1.

ROH patterns in Inner Mongolia cashmere goats. (A) The count of ROH segments exceeding 1 Mb in length per chromosome and distribution of different ROH classes per chromosome (stacked bars). (B) Average percentage of each chromosome covered by ROH (red dots). ROH, runs of homozygosity.

Regarding length distribution (Table 1), short ROHs (1–5 Mb) predominated in number (50.19%), but their cumulative genomic proportion (22.35%) was lower than that of the 5–10 Mb (27.44%) and 10–20 Mb (28.21%) categories. Notably, although ROHs longer than 20 Mb accounted for only 5.36% of the total number, their cumulative length proportion reached 22.01%, which was comparable to that of the short ROH category.

Table 1.

Descriptive statistics of four classes of ROH in IMCG.

Class	N	Percentage of number, %	Mean ± SD, Mb	Total length, Mb	Percentage of length, %
ROH1–5 Mb	1,818	50.19%	3.34 ± 0.84	6070.26	22.35%
ROH5–10 Mb	1,056	29.16%	7.06 ± 1.41	7452.72	27.44%
ROH10–20 Mb	554	15.30%	12.83 ± 2.70	7663.09	28.21%
ROH>20 Mb	194	5.36%	30.82 ± 11.98	5978.23	22.01%
ROHAll	3,622	100.00%	7.50 ± 7.33	27164.29	100.00%

3.2. Genomic inbreeding

Based on genomic data, two inbreeding coefficients were calculated, and the results are presented in Table 2 and Figure 2. The F_HOM ranged from −0.0583 to 0.1654 (mean ± SD: −0.0126 ± 0.0301), with most values clustering between −0.05 and 0.025. In contrast, F_ROH ranged from 0.0009 to 0.1991 (mean ± SD: 0.0290 ± 0.0289) and were mainly distributed between 0 and 0.05. The mean F_ROH values calculated for ROH length categories increased with segment length: 0.0066 ± 0.0037 (1–5 Mb), 0.0091 ± 0.0067 (5–10 Mb), 0.0129 ± 0.0094 (10–20 Mb), and 0.0228 ± 0.0234 (>20 Mb).

Table 2.

Descriptive statistics of inbreeding coefficients for different classes in IMCG.

Inbreeding coefficients	Mean	SD	Range
FHOM	−0.0126	0.0301	−0.0583 to 0.1654
FROH	0.0290	0.0289	0.0009–0.1991
FROH (1–5)	0.0066	0.0037	0.0007–0.0208
FROH (5–10)	0.0091	0.0067	0.0020–0.0459
FROH (10–20)	0.0129	0.0094	0.0041–0.0633
FROH (>20)	0.0228	0.0234	0.0081–0.1207

Figure 2.

Distribution of inbreeding coefficients based on observed homozygosity (F_HOM) and runs of homozygosity (F_ROH) for Inner Mongolia cashmere goats.

Pairwise correlation analysis among these inbreeding coefficients (Figure 3) revealed the strongest association between F_ROH and F_HOM (r = 0.97). The correlation between F_ROH and the category-specific F_ROH estimates increased with ROH length, a pattern also observed for F_HOM. The weakest correlation was found between F_{ROH (1–5)} and F_{ROH (5–10)} (r = 0.35).

Figure 3.

The pearson correlation between the genomic inbreeding coefficients, including F_HOM and F_ROH (F_ROH, F_{ROH(1–5 Mb)}, F_{ROH(5–10 Mb)}, F_{ROH(10–20 Mb)}, and F_{ROH(>20 Mb)}).

3.3. Detection of ROH islands and gene annotation

To identify genomic regions with high-frequency homozygosity in the population, this study focused on SNPs enriched within ROH islands. Figure 4 shows the distribution of the proportion of SNPs located within ROH along the chromosomes, highlighting regions exceeding a predefined threshold. Based on this, a total of 14 ROH islands were identified across 11 chromosomes (Table 3), encompassing 589 SNPs and 238 genes (Supplementary Table S1). Among these, chromosome 25 had the highest number of annotated genes (117 in total), while no genes were annotated in specific regions of chromosomes 7, 8, 13, and 17. Many of the annotated genes are putatively associated with economically important traits, including those related to reproduction (e.g., TXLNA, HDAC1, AQP4) and hair follicle development (e.g., MIR21, DNASE1L2, MMP25).

Figure 4.

Manhattan plot of SNPs frequencies in ROH across the genome. The red line represents the ROH islands threshold for the top 1% of SNP frequency. SNPs, single nucleotide polymorphisms; ROH, runs of homozygosity.

Table 3.

Statistics of ROH islands observed in the IMCG population.

Chromosome	Start, bp	End, bp	Length, bp	No. SNPs	No. genes
2	12,283,827	14,789,346	2,505,520	59	37
2	93,379,497	93,401,647	22,151	2	0
5	49,134,334	54,355,955	5,221,622	121	9
7	5,267,872	7,767,750	2,499,879	75	0
8	11,246,682	13,775,887	2,529,206	47	4
8	105,880,550	105,965,786	85,237	2	0
9	76,818,267	82,654,052	5,835,786	102	27
13	32,488,843	32,550,624	61,782	2	0
17	51,205,345	51,581,703	376,359	10	0
19	10,321,740	13,555,481	3,233,742	63	28
24	30,144,518	30,521,754	377,237	6	1
25	702,894	4,138,138	3,435,245	62	117
26	30,200,702	30,978,147	777,446	31	14
26	41,866,152	42,063,807	197,656	7	1

Further functional enrichment analysis of the genes located within the ROH islands revealed significant results (Figures 5, 6): a total of 16 GO terms related to biological processes, 14 related to molecular functions, and 25 related to cellular components, along with 7 KEGG pathways, were significantly enriched (p < 0.05). Detailed information on the corresponding genes and functional terms is available in Supplementary Table S2. In addition, PPI network analysis revealed that these annotated genes were clustered into nine networks (Figure 7). The largest PPI network contained 142 genes, among which several genes (e.g., RPL3L, RPS2, PTEN, TBL3, GTF2H5, and MRM1) exhibited the highest node degrees.

Figure 5.

GO terms enrichment analysis of genes located within ROH islands in the Inner Mongolia cashmere goats.

Figure 6.

KEGG pathways enrichment analysis of genes located within ROH islands in the Inner Mongolia cashmere goats.

Figure 7.

Protein–protein interaction (PPI) network of candidate genes located within ROH islands in Inner Mongolia cashmere goats.

4. Discussion

This study focused on the IMCG (Erlangshan subtype) due to its well-defined breeding history and accessibility. We acknowledge that the three IMCG subtypes (Albas, Erlangshan, Alashan) may exhibit some genetic differentiation due to geographical separation and breeding objectives (1). Therefore, the ROH patterns and inbreeding coefficients reported here are most directly representative of the IMCG (Erlangshan subtype) population from this specific farm. Future studies encompassing all three subtypes from multiple farms are warranted to fully capture the genomic diversity of the broader IMCG genetic resource.

4.1. Distribution characteristics of runs of homozygosity

A primary consideration in current ROH research is the lack of unified standards (6). The number and length of ROH identified in genotyping data largely depend on the specific parameters and thresholds applied during sequence analysis (19). Additionally, filtering SNPs prior to analysis (e.g., excluding loci with low MAF, deviation from HWE, or high LD) also significantly influences the result (20, 21). It should be noted that adopting stricter criteria, such as increasing the minimum length or reducing the number of allowed heterozygous sites, will reduce both the number of detected ROH and their total length, thereby lowering the F_ROH estimate. This effect is particularly pronounced in short ROH categories, which reflect ancient inbreeding events (13). Conversely, overly permissive parameters may increase the detection rate of false positive ROH (22). We fully recognize that parameter selection directly influences ROH detection outcomes and, consequently, F_ROH estimation. The parameter combination employed in this study represents a widely adopted balanced strategy in livestock genomics research, which was determined after careful consideration in the context of medium-density goat SNP70K BeadChip data.

Using this approach, the mean number of ROH per individual (MN_ROH) detected in Inner Mongolia cashmere goats was 9.35, which closely aligns with the value reported for Chinese Merino (~9.23) (23), yet differs from earlier findings on the same breed by Wang et al. (1). This discrepancy highlights the influence of factors such as SNP density and analytical parameters on ROH detection outcomes (24). Analysis of ROH lengths revealed that fragments between 1 and 5 Mb constituted approximately half (50.19%) of the total, indicating the population has not recently undergone severe inbreeding. This finding aligns with reports by Zhao et al. (25). Notably, ROH longer than 20 Mb still constituted 5.36% of the total. Given that long ROH are associated with an increased risk of deleterious homozygous variants (26, 27) and can lead to harms related to inbreeding such as loss of genetic diversity and deterioration of production traits (28–31), individuals carrying such segments warrant attention due to their potential inbreeding risk.

4.2. Estimation and comparison of inbreeding coefficients

Traditionally, inbreeding coefficients at both the individual and population levels have been primarily calculated from pedigree data (F_PED) using the path coefficient method proposed by Wright (32). This method relies on two key assumptions: pedigree records must be complete and accurate, and individuals within the base population must be unrelated. In practice, however, these conditions are often difficult to satisfy, and pedigree errors due to misunderstandings, misidentification, or recording inaccuracies are relatively common (33). Studies have confirmed that incomplete pedigree information or registration errors can negatively affect the genetic progress of livestock populations and the effectiveness of conservation strategies (34–36).

With the widespread adoption of high-throughput genotyping technologies, the availability of extensive genome-wide SNP data has provided new research tools for various livestock species, including goats. Utilizing this genomic information enables a more precise estimation of inbreeding coefficients, thereby enhancing our understanding of inbreeding. Unfortunately, due to the unavailability of complete pedigree data in this study, we calculated and compared F_ROH and F_HOM in IMCGs. Notably, F_ROH values were generally higher than F_HOM values, and the mean F_HOM was negative (−0.0126). This discrepancy can be attributed to F_HOM is calculated based on the observed proportion of homozygous genotypes across all loci, which cannot distinguish between homozygosity due to identity by descent (IBD) and identity by state (IBS). Consequently, F_HOM can be biased downward, even yielding negative values, when there is an excess of heterozygosity in the population relative to Hardy–Weinberg expectations—a phenomenon that can arise from factors such as balancing selection or population substructure (37, 38). In contrast, F_ROH specifically captures long, continuous homozygous segments that are highly indicative of recent autozygosity (IBD). Therefore, F_ROH is considered a more accurate and direct measure of genomic inbreeding for population management (3). Our observation of higher F_ROH relative to F_HOM is consistent with findings in Arbas Cashmere goats and Hu sheep studies (37, 39).

Furthermore, we analyzed the correlations between inbreeding coefficients based on ROH of different lengths with F_ROH and F_HOM. The results showed a strong correlation between F_ROH and F_HOM (r = 0.97), which is consistent with the findings of Islam et al. (39). Similarly strong correlations have also been observed in other livestock species, such as local cattle (40), buffalo (41), and pigs (42, 43). In this study, the negative values observed for F_HOM may be attributed to heterozygote advantage at certain loci within the population, which increases heterozygote frequency and leads to a lower observed homozygosity than expected (37). In contrast, F_ROH is based on continuous homozygous segments in the genome, and its estimates are always positive. Therefore, F_ROH is considered a more efficient and accurate method for quantifying the level of inbreeding (3, 44).

4.3. Candidate genes within ROH islands and their potential functions

ROH islands refer to continuous homozygous regions in the genome that occur at high frequencies within a population (45). The formation of such regions may result from natural or artificial selection acting on specific genotypes, or from population historical events (e.g., bottleneck effects, genetic drift) or structured breeding practices (3, 22). Therefore, while ROH islands are often interpreted as potential signatures of selection, they are not definitive proof, as identical patterns can arise from neutral processes. Analyzing ROH islands aids in identifying genomic segments that have undergone natural or artificial selection and serves as an important preliminary approach for detecting candidate genes associated with complex traits (46). Consequently, this analytical method has been widely applied across various species to identify genes related to domestication or to animal production and reproductive traits.

Among the 14 ROH islands identified in this study, no annotated genes were detected in 4 islands, which may be attributed to incomplete annotation of the reference genome (47). In the remaining 10 islands, we identified a large number of candidate genes associated with various production traits in IMCGs (Supplementary Table S1). Among them, several genes are considered strong candidates for influencing reproductive processes based on prior functional studies: PUM1 has been shown to regulate the differentiation of embryonic stem cells into oocytes in a mouse model (48); PAQR4, as a member of the PAQR family, may be involved in the progesterone secretion pathway according to a study in dairy goats (49); ABCC2 was identified as an important regulatory factor in the transition from primordial to primary follicles in goats, with its expression declining as follicular development progresses (50); previous studies in other species suggest that HDAC1 and HDAC3 can cooperate with the transcription factor YY1 to regulate PHLDA2, a gene implicated in placental development (51). This highlights HDAC1 as a putative candidate influencing reproductive traits in goats, although direct functional validation in IMCG is needed. Additionally, NDUFB6 and NDUFB10 may contribute to maintaining sperm viability and metabolism (52), while AQP4 and TXLNA could be key molecules affecting reproductive efficiency in goats (53, 54). Notably, MIR21 is not only involved in spermatogenesis and follicular development but also serves as a critical factor in the formation and development of hair follicles in goats (47, 55, 56).

This study also identified multiple candidate genes for hair follicle development: DNASE1L2 has been shown to play an important role in hair follicle development in Changthangi goats and is involved in wool growth regulation and fiber diameter modulation in Gansu alpine fine-wool sheep (57, 58); MMP25 has been determined as a key gene promoting cashmere growth, with its expression significantly upregulated during the active growth phase of cashmere (59). Additionally, LCK, IL32, MEFV, and NLRC3 are often associated with immune regulation (60–63); VMP1, CREBBP, and CHUK are regarded as key regulators of growth and development (64–66); while milk fat synthesis and metabolism are frequently linked to genes such as ACACA, ABCA3, AMDHD2, ECI1, and PGP (67, 68).

In the present study, we further conducted GO term and KEGG pathway analyses for all candidate genes located within ROH islands. Among them, three GO terms are noteworthy: animal organ morphogenesis (GO:0009887), macromolecule metabolic process (GO:0008152), and metabolic process (GO:0043170). Furthermore, accumulating evidence has implicated the mTOR signaling pathway in maintaining hair follicle cycling and follicular stem cell homeostasis (69). These findings collectively suggest that most quantitative phenotypic traits are likely regulated through polygenic interactions. PPI network analysis further highlighted hub genes such as RPL3L, RPS2, PTEN, and TBL3 as central nodes within the interaction network (Figure 7), underscoring their potential regulatory roles in traits like hair follicle development and energy metabolism. Notably, RPS2 participates in a conserved regulatory network responsible for 40S ribosomal subunit assembly (70). RPL3L is a striated muscle-specific ribosomal protein (71), and TBL3 is a confirmed component of the SSU processome essential for small ribosomal subunit biogenesis (72, 73). Furthermore, PTEN is a negative regulator of the mTOR signaling pathway, and its role in hair follicle development has been demonstrated (74). The results provide novel insights into phenotypic trait selection and genetic architecture underlying complex traits. However, it should be noted that the functional information derived from GO and KEGG analyses remains limited in scope. The incomplete annotation of the goat genome and the physical clustering of genes within ROH islands may bias the functional interpretation (38, 46). Future functional validation studies, including tissue-specific expression profiling of key candidate genes (e.g., RPS2, RPL3L, PTEN) in relevant tissues such as skin and hair follicles, are needed to confirm their roles in cashmere production and metabolic traits.

5. Conclusion

In this study, we investigated the distribution of ROH across autosomes and inbreeding coefficients in IMCG (Erlangshan subtype). The ROH segments spanning 1–5 Mb exhibited both the highest frequency and the most substantial cumulative length proportion. Notably, ROH fragments exceeding 20 Mb accounted for 22.01% of the total ROH length, demonstrating a comparable contribution to genomic coverage as the 1–5 Mb category. Correlation analyses among different inbreeding estimators revealed the strongest association between F_ROH and F_HOM. The IMCG population displayed relatively low inbreeding coefficients, suggesting effective management of inbreeding depression. ROH islands were found to encompass multiple candidate genes associated with crucial biological functions including reproductive regulation, hair follicle morphogenesis and development, growth traits, and milk fat metabolism. These findings provide valuable insights for implementing genomic selection strategies in IMCGs while maintaining controlled inbreeding levels. Furthermore, this research establishes a foundation for elucidating the molecular mechanisms underlying economically important traits, thereby facilitating precision breeding programs through targeted utilization of homozygosity patterns.

Funding Statement

The author(s) declared that financial support was received for this work and/or its publication. This reported work was supported by the Science and Technology Program of the Inner Mongolia Autonomous Region (2025YFHH0226), the Fundamental Research Funds of Directly Subordinate Universities of Inner Mongolia Autonomous Region (BR251201), the Program for Innovative Research Team in Universities of Inner Mongolia Autonomous Region (NMGIRT2322), the Inner Mongolia Education Department Special Research Project For First Class Disciplines (YLXKZXNND-007), the Inner Mongolia Autonomous Region Breeding Joint Research Project (YZ2023011), the Science and Technology Plan of Inner Mongolia Autonomous Region (2023KYPT0021), and the Project of Northern Agriculture and Livestock Husbandry Technical Innovation Center, Chinese Academy of Agricultural Sciences (BFGJ2022002).

Data availability statement

The datasets presented in this study can be found in online repositories. The datasets generated for this study are available in the Figshare repository under DOI: https://doi.org/10.6084/m9.figshare.29224259.

Ethics statement

The animal studies were approved by Scientific Research and Academic Ethics Committee of Inner Mongolia Agricultural University and the Biomedical Research Ethics of Inner Mongolia Agricultural University (Approval No. [2020] 056). The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent was obtained from the owners for the participation of their animals in this study.

Author contributions

YQ: Methodology, Software, Formal analysis, Writing – original draft. YR: Formal analysis, Validation, Writing – original draft. MZ: Data curation, Investigation, Writing – original draft. XW: Formal analysis, Visualization, Writing – original draft. XA: Data curation, Writing – original draft. FS: Formal analysis, Writing – original draft. YZ: Funding acquisition, Supervision, Writing – review & editing. RW: Conceptualization, Funding acquisition, Methodology, Project administration, Writing – review & editing.

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that Generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fvets.2026.1796462/full#supplementary-material