Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2024 Nov 27;142(4):408–418. doi: 10.1111/jbg.12914

Genomic Diversity of U.S. Katahdin Hair Sheep

Sara M Nilson 1,✉, Joan M Burke 2, Gabrielle M Becker 3, Brenda M Murdoch 3, Jessica L Petersen 1, Ronald M Lewis 1
PMCID: PMC12149482  PMID: 39602081

ABSTRACT

In the late 1950s, Katahdin hair sheep were developed as a composite breed of medium size and moderate prolificacy, with potential to express resistance to gastrointestinal nematodes. With increasing popularity and the recent adoption of genomic prediction in their genetic evaluation, there is a risk of decreasing variation with selection based on genomically enhanced estimated breeding values. While Katahdin pedigrees are readily available for monitoring diversity, they may not capture the entirety of genetic relationships. We aimed to characterise the genomic population structure and diversity present in the breed, and how these impact the size of a reference population necessary to achieve accurate genomic predictions. Genotypes of Katahdin sheep from 81 member flocks in the National Sheep Improvement Program (NSIP) were used. After quality control, there were 9704 animals and 31,984 autosomal single nucleotide polymorphisms analysed. Population structure was minimal as a single ancestral population explained 99.9% of the genetic variation among animals. The current N e was estimated to be 150, and despite differences in trait heritabilities, the effect of N e on the accuracy of genomic predictions suggested the breed should aim for a reference population size of 15,000 individuals. The average degree of inbreeding estimated from runs of homozygosity (ROH) was 16.6% ± 4.7. Four genomic regions of interest, previously associated with production traits, contained ROH shared among > 50% of the breed. Based on four additional methods, average genomic inbreeding coefficients ranged from 0.011 to 0.012. The current population structure and diversity of the breed reflects genetic connectedness across flocks due to the sharing of animals. Shared regions of ROH should be further explored for incorporation of functional effects into genomic predictions to increase selection gains. Negative impacts on genetic diversity due to genomic selection are not of immediate concern for Katahdin sheep engaged in NSIP.

1. Introduction

Since 1942, the United States (U.S.) sheep industry has dramatically decreased in size from 56 million to 5 million sheep (National Research Council et al. 2008; USDA National Agricultural Statistics Service, 2020). The relatively quick decrease in size may have reduced the amount of genetic variability within the population resulting in the need to diversify by crossbreeding to adapt to the changing landscape of the industry (Falconer 1960; Thorne et al. 2021). In addition, sheep populations have shifted geographically with an increase in numbers in the upper midwest and eastern U.S. accompanied by a proportional decrease in Texas and New Mexico (Thorne et al. 2021). Climatological changes will necessitate adaptability to maintain productivity in the face of new challenges, such as greater exposure to parasites (McKenzie and Williams 2015). One appealing avenue for producers may be crossbreeding or the utilisation of a composite breed that could capitalise on the beneficial traits of multiple breeds while potentially increasing the amount of genomic diversity available for selection.

The Katahdin, a composite breed created in the late 1950s in the northeastern U.S. is medium‐sized and moderately prolific, with the potential to show resistance to gastrointestinal nematodes (Wildeus 1997; Notter 2000). The breed combines the freedom from wool through shedding, thereby reducing production costs, with ample fitness, muscle, and growth (Notter 2000). Due to the combined benefit of these breed attributes, the number of registered Katahdin sheep has increased nationally with their enrolment in the National Sheep Improvement Program (NSIP) steadily rising (Notter 2000; Thorne et al. 2021). As of August 2020, the Katahdin breed accounted for the largest proportion, 26.9%, of enrolled sheep in the NSIP (Thorne et al. 2021). In the U.S. the NSIP serves as a flexible programme for data consolidation and provides purebred and commercial sheep producers with comprehensive across‐flock genetic evaluations through the reporting of estimated breeding values (EBV) and selection indexes (Wilson and Morrical 1991; Notter 1998).

Pedigrees characterise relationships among animals and as demonstrated through time, reliably guide producers' breeding decisions. Pedigrees provide an estimate of the extent of shared ancestry given probabilities of genetic relatedness (e.g., full siblings share 50% of their genome) (Wright 1922). The extent of genetic diversity in Katahdin sheep enrolled in NSIP has recently been assessed based on their extensive pedigree information (Nilson et al. 2024). Briefly, inbreeding was low (1.62%) and the rate of increase in inbreeding was < 1% per generation. The average effective population size (N e ) was estimated to be 111 animals which was above the FAO level of concern (N e  = 50) to avoid rapid loss of genetic diversity (FAO Animal Production and Health Division, 1998). With a stable low level of inbreeding, conservation of variation was not found to be an immediate concern based upon that pedigree‐based analysis of the NSIP Katahdin population (Nilson et al. 2024).

Pedigrees, however, do not account for differences in kinship due to Mendelian sampling. In addition, pedigrees are prone to error through the misidentification of and missing ancestors. Misidentification of animals or relationships is especially challenging in the sheep industry due to several factors including that sheep are a litter bearing species and ewe behaviour of ‘lamb stealing’ (Alexander, Stevens, and Mottershead 1983; Gonyou and Stookey 1985; Thorne et al. 2021). Meanwhile, advancements in genomics have allowed animal relationships to be defined at a higher resolution that can assist in resolving common pedigree errors (Wang 2016).

Recently, NSIP has adopted the use of genomic prediction in Katahdin sheep (McMillan et al. 2022). Since the use of genomically enhanced estimated breeding values (GEBV) in selection decisions may increase the risk of losing genetic variation, diversity in this breed should be monitored at a finer scale (Wray and Goddard 1994). The genomic diversity of the Katahdin breed is of great interest for progressive NSIP producers who want to improve the productivity and well‐being of their sheep in the face of global challenges. This research aimed to (1) characterise the genomic population structure and diversity present in the breed, and (2) determine the subsequent impact on the accuracy of genomic predictions. Characterising the genomic diversity present in the population will assist NSIP producers with their breeding programmes by clarifying animal relationships, which will accelerate genetic gains for traits of interest.

2. Methods

2.1. IACUC Statement

The data utilised in this study were previously collected by the animals' owners in commercial Katahdin flocks that were members of NSIP and, therefore, did not require institutional animal care and use approval.

2.2. Genomic Data

The NSIP provided raw genotypes for 9775 sheep with birth years spanning from 2007 to 2023 in the Katahdin pedigree. Genotypes for individuals were generated on one of two arrays: the GeneSeek Genomic Profiler (GGP) Ovine 50K BeadChip versions B (n = 4482) or C1 (n = 5293). Single nucleotide polymorphisms (SNP) from all arrays were remapped to the ARS‐UI_Ramb_v2.0 reference using the National Center for Biotechnology Information Genome Remapping Service (Davenport et al. 2022). Only known shared autosomal SNP positions were kept. In Plink v1.9 (Purcell et al. 2007; Chang et al. 2015), SNP with duplicated positions were reduced to those with the highest call rate. For quality control, SNP with a call rate < 90% (n = 656) and animals with a genotyping rate < 90% (n = 71) were removed. As minor allele frequency and linkage disequilibrium filtering can alter the detection of population structure and runs of homozygosity (ROH), no more SNP filtering was applied for these analyses (Linck and Battey 2019; Meyermans et al. 2020). However, a minor allele frequency threshold of 0.001 was applied before evaluating inbreeding and heterozygosity to remove fixed loci.

2.3. Population Structure

Population structure was initially examined through principal component analysis (PCA) and identity by state (IBS) clustering with the ‘‐‐pca’ and ‘‐‐cluster’ flags, respectively, in Plink v1.9 (Purcell et al. 2007; Chang et al. 2015). For PCA, the first 100 principal components were estimated. No prior information (e.g., flock identifier) was provided for population structure analyses. If population structure was indicated, genetic differentiation was formally tested between populations with the ‘‐‐fst’ flag in Plink v1.9 (Purcell et al. 2007; Chang et al. 2015).

In addition, the number of populations, K, contributing to the current variation was estimated using fastStructure v1.0 (Raj, Stephens, and Pritchard 2014). A range of K values from 1 to 20 were considered. Two statistics to help determine the ‘true’ K were obtained: the K that explained 99.9% of the variation and the K that maximised the log‐marginal likelihood (Raj, Stephens, and Pritchard 2014).

2.4. Effective Population Size

Effective population size was estimated from linkage disequilibrium using SNeP v1.11 (Barbato et al. 2015). Default parameters were used, except that bins with more than 75 pairwise comparisons were kept. A linear model was fitted to the SNeP v1.11 estimations to predict N e for more recent generations as the Katahdin breed is relatively new.

The effects of N e on the accuracy of genomic prediction were estimated using four different methods (Daetwyler, Villanueva, and Woolliams 2008; Goddard 2009; Daetwyler et al. 2010; Goddard, Hayes, and Meuwissen 2011; Karaman et al. 2016). These accuracies can serve as baseline estimates for animals with no direct relatives in the reference population and no phenotype for the trait of interest. As these accuracies are sensitive to the assumptions unique to each method, they were utilised to investigate the size of the reference population, namely the number of animals with genotypes and phenotypes, that may correspond to a plateau in prediction accuracy (Brard and Ricard 2015). This inflection point associated with the plateau in prediction accuracy would be a target number of animals the Katahdin breed should include in the reference population given their N e . The reference population size explored ranged from 1k to 25k for heritabilities of 0.1, 0.3, and 0.5. The number of chromosomal segments was estimated by 2N e L where L was the length of the genome in morgans (26 M) (van der Werf 2013; Davenport et al. 2022). When the number of SNP to be utilised was included in a method, we used 45,000 corresponding to ~90% of SNP passing quality control from a 50K BeadChip, which is the most utilised commercial genotyping platform for US sheep.

2.5. Runs of Homozygosity

Runs of homozygosity were detected in Plink v1.9 with the ‘‐‐homozyg’ flag (Purcell et al. 2007; Chang et al. 2015). Scanning windows consisted of 15 SNP allowing a maximum of two heterozygous and one missing SNP (Becker et al. 2024). The minimum ROH length was set to 250 kb and contained at minimum 30 SNP. Runs had a minimum density of 1 SNP per 10 kb with a maximum internal gap of 10 kb.

Individual inbreeding coefficients (F ROH) were estimated as the proportion of the genome coverage by ROH. Genome length was estimated by the amount of genome covered by ROH in a pseudo‐animal that was homozygous for every SNP. Average per cent coverage of chromosomes by ROH across the population was estimated following the method proposed by Al‐Mamun et al. (2015). Length of the chromosome was defined as the cumulative length covered by the ROH detected in the pseudo‐animal. The pseudo‐animal was simulated to estimate genome and chromosomal length to consider potential undetected gaps in coverage due to SNP density.

Regions with overlapping ROH shared among more than 50% of the population were identified as ROH islands. Of interest, these regions were explored for associations with previously identified quantitative trait loci through query of the Sheep Quantitative Trait Locus (QTL) Database (db) using the Animal QTLdb (Hu, Fritz, and Reecy 2007).

2.6. Linkage Disequilibrium Pruning

For non‐ROH inbreeding and heterozygosity estimates only, linkage disequilibrium pruning was performed using two methods. The first method, implemented in Plink v1.9 with the flag ‘‐‐indep‐pairwise’, used a pruning threshold of r 2  = 0.8, a window size of 50 kb, and a step size of 5 variants (Purcell et al. 2007; Chang et al. 2015). Window size was chosen due to the average gap size of 50.7 kb between variants on the GGP Ovine 50 K BeadChip.

The second method used a priori knowledge of the rate of linkage disequilibrium decay in sheep breeds while considering SNP density. On average across breeds, linkage disequilibrium reached an r 2 = 0.2 at approximately 30 kb between SNP for the Illumina OvineSNP50 BeadChip (Kijas et al. 2014). Implemented in Plink v1.9 with the ‘‐‐bp‐space’ flag, a single SNP was randomly selected within a window size of either 20 or 30 kb (Purcell et al. 2007; Chang et al. 2015). The inbreeding and heterozygosity results from the three pruning methods were compared to determine if a priori biological knowledge can be utilised instead of threshold self‐derived data to achieve accurate estimations.

2.7. Inbreeding and Heterozygosity

Inbreeding was estimated with four methods in Plink v1.9.l. Specifically, they were the method‐of‐moments (F MM) with the ‘‐‐het’ flag and with the ‘‐‐ibc’ flag, the variance‐standardised relationship minus one (F VStRel‐1), the excess of homozygosity (F EH), and the correlation of uniting gametes (F UG) (Purcell et al. 2007; Chang et al. 2015). For comparison, pedigree‐based inbreeding coefficients (F PED) were estimated using ENDOG v4.8 (Gutiérrez and Goyache 2005; Nilson et al. 2024). Observed and expected heterozygosities were estimated as the difference from homozygosities in Plink v1.9 with the ‘‐‐het’ flag (Purcell et al. 2007; Chang et al. 2015). Spearman and Pearson correlations were estimated among inbreeding coefficient estimates across methods to estimate the trend and rank differences.

3. Results

3.1. Genomic Data

After initial quality control, 9704 animals (4081 male and 5623 female) and 31,984 SNP were available for analysis. This represented 9.7% of animals with a body weight recorded. There were 81 flocks ranging in size from 1 to 1392 with an average of 120 animals. This full SNP set was utilised for population structure, N e , and ROH analyses. For the inbreeding and heterozygosity analyses, after the minor allele frequency filter of 0.001 was applied, 31,859 SNP remained. When linkage disequilibrium pruning was applied with the threshold of r 2  = 0.8, there were 2014 SNP removed with 29,845 remaining for analysis. When linkage disequilibrium pruning was applied with the random selection of a SNP every 20 kb, there were 5511 SNP removed with 26,348 remaining for analysis. When linkage disequilibrium pruning was applied with the random selection of a SNP every 30 kb, there were 7597 SNP removed with 24,262 remaining for analysis.

3.2. Population Structure

From the PCA, principal components one and two explained 4.9% and 3.8% of the genetic variation in the population (Figure 1). A single flock trailed away from the rest of the population on PC2 as it decreased below −0.02. Genetic differentiation between the rest of the population and that flock was formally tested with a F ST = 0.032 ± 0.041.

FIGURE 1.

FIGURE 1

Open in a new tab

Principal component analysis. Principal components one (PC1) and two (PC2) explain 4.9% and 3.8% of the genetic variation among the animals, respectively. [Colour figure can be viewed at wileyonlinelibrary.com]

A single ancestral population (K = 1) explained 99.9% of the current population variation based on fastStructure, while a K = 15 maximised the log‐marginal likelihood. As K increased, family lines rather than flocks began to become distinguishable (Figure S1).

3.3. Effective Population Size

Overall, the estimated N e has decreased since the founding of the breed (N e  = 315, ~22 generations ago) and is currently (in 2023) estimated to consist of 150 animals. Based on the N e of 150, the estimated number of effective chromosomal segments in the population was 7800. Using the breed's genetic diversity as defined by the number of effective chromosomal segments, the accuracies of genomic predictions were estimated for a reference population size ranging from 1k to 25k animals. Across methods, prediction accuracies began to plateau when the reference population consisted of 15k animals (Figure 2).

FIGURE 2.

FIGURE 2

Open in a new tab

Effective population size effects on accuracy of genomic predictions. The accuracy of genomic predictions for heritabilities of 0.1 (blue, solid), 0.3 (purple, short dash), and 0.5 (orange, long dash) for the estimated effective population size of 150. Methods investigated were: (a) Goddard (2009), (b) Goddard, Hayes, and Meuwissen (2011), (c) Daetwyler, Villanueva, and Woolliams (2008); Daetwyler et al. (2010) and (d) Karaman et al. (2016). [Colour figure can be viewed at wileyonlinelibrary.com]

3.4. Runs of Homozygosity

The total number of ROH detected was 713,157 with an average of 73.5 runs per animal. The range of ROH length was 0.25–179.5 Mb with an average of 5.56 Mb and a median of 4.01 Mb (Table 1).

TABLE 1.

Length distribution of runs of homozygosity.

Length Runs of homozygosity (n)
Mb ≤ 1 21,420
1 < Mb ≤ 5 422,997
5 < Mb ≤ 10 188,027
10 < Mb ≤ 15 48,537
15 < Mb ≤ 20 16,969
20 < Mb ≤ 25 7075
25 < Mb 8132
Open in a new tab

Note: The distribution of the number of runs of homozygosity by length in megabases (Mb). Runs of homozygosity length ranged from 0.25 to 179.5 Mb with an average of 5.56 Mb and a median of 4.01 Mb.

The estimated F ROH ranged from 0.37% to 51.2% with an average of 16.6% ± 4.7%. Pearson and Spearman correlations were high between F ROH and the other genomic inbreeding coefficients except for the coefficient derived from the F VStRel‐1 (Tables 2 and 3; Table S1). In addition, correlations among F ROH and the other genomic inbreeding coefficients were similar across LD pruning methods (Tables 2 and 3; Table S1). The average per cent coverage of autosomes was 18.4% and ranged from 15.1% (Chromosome 9) to 27.1% (Chromosome 23) (Figure 3).

TABLE 2.

Pearson and Spearman correlations among inbreeding estimates from systematic linkage disequilibrium pruned data.

Inbreeding statistic F MM F VStRel‐1 F EH F UG F ROH F PED
F MM −0.05 0.90 0.72 0.98 0.57
F VStRel‐1 −0.04 −0.46 0.65 −0.05 0.17
F EH 0.92 −0.32 0.38 0.89 0.44
F UG 0.72 0.57 0.49 0.70 0.55
F ROH 0.97 −0.05 0.89 0.68 0.57
F PED 0.49 −0.02 0.49 0.36 0.49
Open in a new tab

Note: The Pearson (upper triangle) and Spearman (lower triangle) correlations of genomic inbreeding from the systematically linkage disequilibrium pruned data (r 2  = 0.8). The inbreeding coefficients were estimated by method of moments (F MM), the variance standardised relationship—1 (F VStRel‐1), excess homozygosity (F EH), and the correlation of uniting gametes (F UG).

TABLE 3.

Pearson and Spearman correlations among inbreeding estimates from random 20 kb linkage disequilibrium pruned data.

Inbreeding statistic F MM F VStRel‐1 F EH F UG F ROH F PED
F MM −0.02 0.90 0.73 0.98 0.57
F VStRel‐1 −0.01 −0.42 0.66 −0.04 0.18
F EH 0.92 −0.29 0.40 0.90 0.45
F UG 0.72 0.58 0.50 0.71 0.55
F ROH 0.97 −0.03 0.90 0.68 0.57
F PED 0.49 −0.02 0.49 0.35 0.49
Open in a new tab

Note: The Pearson (upper triangle) and Spearman (lower triangle) correlations of genomic inbreeding from the random linkage disequilibrium pruned data (20 kb). The inbreeding coefficients were estimated by method of moments (F MM), the variance standardised relationship—1 (F VStRel‐1), excess homozygosity (F EH), and the correlation of uniting gametes (F UG).

FIGURE 3.

FIGURE 3

Open in a new tab

Percent coverage of chromosomes by runs of homozygosity. The height of each bar corresponds to the coverage (%) of each chromosome by runs of homozygosity (ROH). The average ROH coverage was 18.4% (green, solid line). [Colour figure can be viewed at wileyonlinelibrary.com]

Four ROH islands were identified on chromosomes 2, 6, 23, and 25 (Figure 4). The core subregions with the highest shared percentages of 56%, 80%, 71%, and 54% were 0.4, 0.2, 0.6, and 0.6 Mb in length for chromosomes 2, 6, 23, and 25, respectively. These subregions were primarily associated with QTL related to body weight, carcass, milk, and wool traits (Tables S2–S5).

FIGURE 4.

FIGURE 4

Open in a new tab

Single nucleotide polymorphism in shared runs of homozygosity. The frequency (%) shared single nucleotide polymorphisms (SNP) appeared in runs of homozygosity. Regions of interest (highlighted in green) were shared among greater than 50% (green, solid line) of the animals. [Colour figure can be viewed at wileyonlinelibrary.com]

3.5. Inbreeding and Heterozygosity

Average genomic inbreeding coefficients did not differ across the methods utilised for linkage disequilibrium pruning (Table 4). In addition, all linkage disequilibrium pruning methods were highly correlated (0.99) with each other within each method of inbreeding coefficient estimation (Table 5). The range of inbreeding coefficients differed among estimation methods (Figure 5; Figures S2 and S3). Pearson and Spearman correlations among inbreeding coefficients were moderate to high with the greatest differences being observed when F VStRel‐1 was compared to F MM, F ROH, or F PED (Tables 2 and 3; Table S1).

TABLE 4.

Means and standard deviations of genomic inbreeding estimates.

Method SNP F MM F VStRel‐1 F EH F UG
Systematic 29,845 0.011 ± 0.053 0.012 ± 0.094 0.012 ± 0.077 0.012 ± 0.047
Random 20 kb 26,348 0.011 ± 0.053 0.012 ± 0.094 0.012 ± 0.077 0.012 ± 0.047
Random 30 kb 24,262 0.011 ± 0.053 0.012 ± 0.094 0.012 ± 0.077 0.012 ± 0.047
Open in a new tab

Note: The means and standard deviations of genomic inbreeding estimates across the three linkage disequilibrium pruning methods: systematic (r 2  = 0.8), random pruning at 20 kb and random pruning at 30 kb. The total number of SNP analysed for each method are included in the SNP column. The inbreeding coefficients were estimated by method of moments (F MM), the variance standardised relationship—1 (F VStRel‐1), excess homozygosity (F EH), and the correlation of uniting gametes (F UG).

TABLE 5.

Pearson and Spearman correlations among genomic inbreeding estimates from linkage disequilibrium pruned data.

Method comparison F MM F VStRel‐1 F EH F UG
Syst. vs Rand. 20 kb 0.998 (0.996) 0.998 (0.995) 0.997 (0.995) 0.998 (0.994)
Syst. vs Rand. 30 kb 0.997 (0.994) 0.997 (0.993) 0.996 (0.992) 0.997 (0.991)
Rand. 20 kb vs Rand. 30 kb 0.999 (0.998) 0.999 (0.998) 0.999 (0.997) 0.999 (0.997)
Open in a new tab

Note: The Pearson and Spearman (within parentheses) correlations of genomic inbreeding estimates across the three linkage disequilibrium pruning methods: systematic (Syst., r 2  = 0.8), random pruning at 20 kb (Rand. 20 kb) and random pruning at 30 kb (Rand. 30 kb). The inbreeding coefficients were estimated by method of moments (F MM), the variance standardised relationship – 1 (F VStRel‐1), excess homozygosity (F EH), and the correlation of uniting gametes (F UG).

FIGURE 5.

FIGURE 5

Open in a new tab

Distribution of estimated inbreeding coefficients. Methods are denoted as F MM for method of moments, F VStRel‐1 for the variance‐standardised relationship minus one, F EH for the excess of homozygosity, F UG for the correlation between uniting gametes, F ROH for runs of homozygosity and F PED for the pedigree estimate. The first four statistics were estimated from the systematically pruned dataset. [Colour figure can be viewed at wileyonlinelibrary.com]

Heterozygosity estimates were significantly less (paired t‐test, p < 2.2 × 10−16) when calculated using the SNP set pruned for LD using an r 2 threshold compared to either random pruning method; there was no significant difference between the random methods (Table 6).

TABLE 6.

Means and standard deviations of genomic heterozygosity estimates.

Method Observed heterozygosity Expected heterozygosity
Systematic 0.379 ± 0.0202 0.383 ± 0.0003
Random 20 kb 0.382 ± 0.0203 0.386 ± 0.0003
Random 30 kb 0.382 ± 0.0203 0.386 ± 0.0003
Open in a new tab

Note: The mean and standard deviations of the observed and expected heterozygosities across the three linkage disequilibrium pruning methods: systematic (r 2  = 0.8), random pruning at 20 kb and random pruning at 30 kb.

4. Discussion

The Katahdin hair breed is an asset for the sheep industry, known not only for their growth and carcass traits but also for their potential resistance to internal parasites. While pedigree diversity of the NSIP Katahdin has been characterised (Nilson et al. 2024), genomic diversity estimates allow for a more precise definition of relationships among animals within the breed (Wang 2016). The increased clarity provided by describing the current genomic diversity will help producers enhance genetic gains while minimising the potential negative side effects of selection.

The structure of a population or a breed can influence genetic diversity as geographically separation or different breeding practices among flocks could limit the dispersal of genetics. Even though they are spread out across the U.S. the 81 Katahdin flocks are genetically connected (Figure 1, Figure S1). While one flock appeared to be branching genetically from the others, this difference may reflect the breeding programme in this flock. Within it, a greater proportion of sires were retained from within the flock (n = 476) than from outside flocks (n = 136), which may have resulted in sampling bias among animals genotyped. Breeding strategies employed by most Katahdin producers instead entails family lines being shared by purchasing of rams and replacement ewes from outside flocks, which would reduce the risk of loss of variation and increased inbreeding due to restricted gene flow (Simianer et al. 2021).

One important measure of genomic diversity is inbreeding or the proportion of homozygous loci in an animal's genome. Breeding two related and highly inbred animals often increases homozygosity among their offspring, with potential negative impacts on their performance (Vostry et al. 2018). Genomic inbreeding in the Katahdin breed on average was low (0.01), and lower than pedigree‐based inbreeding estimates (0.02) (Nilson et al. 2024). Previous genomic estimates with a subsample (n = 4884) of this population were similar (F MM = 0.01), but were an order of magnitude smaller when compared to another subsample (n = 581, F MM = 0.1) (Becker, Burke, and Lewis 2022; Becker et al. 2024). Compared to seven other US breeds F EH (0.05–0.16), the Katahdin had smaller F EH (0.012) (Davenport et al. 2020; Wilson et al. 2022). Lower inbreeding in the Katahdin may reflect conscientious producers utilising pedigrees in their breeding decisions, the sharing of family lines across flocks as seen in the population structure analyses or simply the lower inbreeding may be a result of the Katahdin being a composite breed. Additionally, the breed registry operates as an open flock book by permitting upgrading to full ‘Katahdin’ (at least 87.5%), which allows introgression of genetics from outside breeds offering favourable attributes.

With the goal of potentially releasing a genomic inbreeding estimate to producers for assisting in breeding decisions, we compared how different methods of pruning for LD would affect inbreeding coefficient and heterozygosity estimates. The effects of drift, selection intensity, and N e can result in greater LD and an increase in information redundancy due to tightly linked markers, which can potentially affect genomic predictions (Hill and Robertson 1968; Dementieva et al. 2021). It is common practice to prune SNP when they are in high LD to reduce bias due to multicollinearity (Wiggans et al. 2009; Malomane et al. 2018). Pruning for LD in a pairwise systematic method using an r 2 threshold can decrease the observation of rare alleles creating an apparent excess of common alleles that would artificially change population structure (Dementieva et al. 2021). Furthermore, LD pruning using an r 2 threshold could also be biased depending upon known or cryptic population structure. Here, there was no difference among inbreeding coefficient estimates across LD pruning methods tested despite differences in SNP number. Observed and expected heterozygosities were greater (p < 2.2 × 10−16) using the markers chosen at random as compared to those resulting from systematic pruning, which corresponded with an increase in SNP with a lower minor allele frequency. Given the higher recombination rate in sheep as a species, the random pruning method may be more reflective of the true heterozygosity measure. An expectation of larger differences due to LD pruning method may be offset in this dataset due to the larger sample size and the lack of population structure.

Additionally, we used six methods for estimating genomic inbreeding with the goal of identifying a preferred measure to release to producers. There is observable skewness in the F VStRel‐1 and F EH estimates, but this is expected given the anticipated ranges, [−1, ∞] and [−∞, 1], respectively, of these statistics (Caballero, Villanueva, and Druet 2021; Villanueva et al. 2021; Dadousis et al. 2022). Due to the difficult interpretation of these two coefficients with their large ranges and variances, and inconsistent estimates in the face of directional selection, F VStRel‐1 and F EH are not optimal coefficients for producers to utilise. Meanwhile, the inbreeding coefficient F MM consistently had the highest Pearson and Spearman correlations with F PED and F ROH (Tables 2 and 3; Table S1). While FROH can be scaled to a [0, 1] range, its estimate is highly dependent on the parameters utilised to estimate the number of ROH, which can easily lead to inconsistencies (Caballero, Villanueva, and Druet 2021; Villanueva et al. 2021; Dadousis et al. 2022). However, F MM is bound from [−1, 1] and may be an inbreeding coefficient that producers may better understand; as homozygosity increases, an animal's inbreeding coefficient becomes closer to 1 and as heterozygosity or diversity is gained relative to the population the coefficient approaches −1 (Caballero, Villanueva, and Druet 2021; Villanueva et al. 2021; Dadousis et al. 2022). Additionally, since F MM is moderately correlated with F PED, an advantage of its use is that large changes in inbreeding coefficients between the genomic and pedigree‐based estimates would most likely be due to pedigree errors or missing information. By correcting for missing or mistaken ancestry, this genomic inbreeding estimate would benefit producers in their selection and mating decisions (Lutaaya et al., 2001; Reid et al., 2014). Therefore, F MM seems the preferred genomic inbreeding statistic to release to sheep producers.

Many selection programmes aim to increase the frequency of beneficial alleles for traits of economic importance. These selection pressures unintendedly contribute to decreasing genetic diversity by increasing homozygosity in the surrounding regions. By detecting ROH, the amount of homozygosity present can be quantified as a proportion of the genome. On average, animals had about 16.6% of their genome and 18.4% of their chromosomes covered by ROH. Yet, there were several animals with over 45% of their genome in ROH. While the average F ROH (18.7%) from a subset of individuals in a previous study was higher, the highest estimate of F ROH was smaller (42.8%) (Becker et al. 2024). Regions with potential beneficial effects on traits of interest may therefore be identified and incorporated into breeding programmes.

Four highly conserved ROH islands were shared among 54%–80% of the sampled animals. These regions were associated with previously identified QTL for, primarily, carcass, wool, and milk traits. The QTL on Chromosome 25 were associated with fleece variation in addition to the delineation of hair and wool types (Demars et al. 2017). One of these regions, Chromosome 23, stands out particularly as 27.1% of that chromosome was covered by ROH and was previously associated with levels of immunoglobulin A in response to gastrointestinal parasites (Atlija et al. 2016). These two core regions overlapped with previously identified ROH islands in the Katahdin breed (Becker et al. 2024). In addition, two of these identified regions, Chromosomes 2 and 6, were associated with faecal egg counts (Davies et al. 2006; Silva et al. 2012; Al Kalaldeh et al. 2019). As the Katahdin are recognised for their parasite resistance, these regions hold promise for functional effects that could further improve resistance within the breed or be crossbred into other breeds.

Genomic diversity within the Katahdin breed is reflected in the estimated N e of 150 which is above the threshold of concern for inbreeding (FAO Animal Production and Health Division, 1998). This was greater than the average pedigree‐based estimate of N e of 111 (Nilson et al. 2024) but smaller than a previous genomic estimate from a subset of the population (N e  = 161) (Becker et al. 2024). The genomic N e may be a better representative of the diversity within the breed as it is derived from shared genomic regions among individuals rather than the probability of their ancestral relationships. However, the use of GEBV in selection decisions may decrease N e as elite genetics are identified and incorporated across flocks (Goddard 2009). Popular sires selected because of their desirable GEBV tend to breed sons that become popular sires. This results in more chromosomal segments being shared among the population, reducing variability (van der Werf 2013).

To counteract the potential loss of diversity and reducing N e through selection, Katahdin flocks should continue to maintain multiple shared family lines, which can be achieved through the purchase of rams or replacement ewes. By adopting breeding strategies like optimal contribution selection, producers can manage relatedness among individuals while maximising genetic gains (Wray and Goddard 1994; Meuwissen 1997). In addition, the sharing of family lines across flocks should be encouraged as genomic connectedness will assist in the reduction of bias in genomic prediction as environmental and genetic effects will be easier to separate (Kennedy and Trus 1993; Yu et al. 2017). Based on our analyses, a reference population size of 15k would optimise the trade‐off between improving accuracy and genotyping costs relative to animal value in NSIP Katahdin sheep. Producers are quickly nearing this goal with nearly 13k animals already genotyped (NSIP, personal communication).

Maintaining diversity in the breed will have a positive impact for Katahdin producers. One economically important trait that Katahdin animals' express is the resistance to gastrointestinal nematode parasites which has increasing value as anthelmintic treatments are becoming ineffective (Burke and Miller 2004; Ngere et al. 2018; Thorne et al. 2021; Arisman et al. 2023). The Katahdin breed has made tremendous gains in parasite resistance and correlated traits like number of lambs weaned, but loss of genetic diversity will erode progress if there is no variation to select upon (Thorne et al. 2021; Rafter et al. 2022). Preserving genetic variation will ensure the breed can maintain an adaptive response to shifting environmental pressures like increased parasite load and continue to improve economically important traits (Taberlet et al. 2011; Thorne et al. 2021; Becker et al. 2024).

5. Conclusion

Katahdin hair sheep breeders have maintained genetic diversity in their sheep while actively engaged in breed improvement through the NSIP. On average, genomic inbreeding coefficients were estimated to be low. As a result of mindful breeding practices with the assistance of pedigrees, inbreeding levels are currently not a concern. However, they should be re‐evaluated in the future to monitor potential loss of diversity with the adoption of genomic selection. Shared homozygous regions of interest were associated with previously determined QTL relating to economically relevant production traits. These regions could be further investigated for functional effects that could be incorporated into genomic predictions to further accelerate selection response. There was little to no population structure as family lines were shared across Katahdin flocks likely through purchase of sires and replacement females. This genetic connectedness among flocks lends itself to reducing bias in genomic predictions as environmental and genetic effects will be more reliably delineated. While the current effective population size of the Katahdin breed suggests as a goal a reference population of 15k, it will need to be continuously refreshed with newly genotyped animals and continued recording of phenotypes to maintain the desired levels of accuracy. As producers use GEBV in their selection decisions, the genomic diversity present in Katahdin sheep is posed to allow substantial genetic gains to be achieved in this breed.

Author Contributions

S.M.N. and R.M.L. conceptualized and designed the work. S.M.N. analysed and interpreted the data and prepared the first draft of the manuscript. R.M.L., J.M.B., and J.L.P. provided guidance and funding. All authors read, revised, and approved the final manuscript.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1.

JBG-142-408-s001.zip (521.2KB, zip)

Table S1.

JBG-142-408-s002.docx (35.8KB, docx)

Acknowledgements

We thank the many U.S. sheep associations, including the Eastern Alliance for Production Katahdins, the Katahdin Hair Sheep International, the National Sheep Improvement Program and Katahdin sheep producers, for their contributions to this research. This work was completed utilising the Holland Computing Center of the University of Nebraska, which receives support from the UNL Office of Research and Economic Development, and the Nebraska Research Initiative. Any opinions, findings, conclusions or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture. Mention of trade names or commercial products in this manuscript is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture. The U.S. Department of Agriculture (USDA) prohibits discrimination in all its programmes and activities on the basis of race, colour, national origin, age, disability and where applicable, sex, marital status, familial status, parental status, religion, sexual orientation, genetic information, political beliefs, reprisal or because all or part of an individual's income is derived from any public assistance programme. (Not all prohibited bases apply to all programmes.) Persons with disabilities who require alternative means for communication of programme information (Braille, large print, audiotape, etc.) should contact USDA's TARGET Center at (202) 720‐2600 (voice and TDD). To file a complaint of discrimination, write to USDA, Director, Office of Civil Rights, 1400 Independence Avenue, S.W., Washington D.C. 20250‐9410, or call (800) 795‐3272 (voice) or (202) 720‐6382 (TDD). USDA is an equal opportunity provider and employer.

Funding: This work was supported by the Organic Agriculture Research and Extension Initiative (grant no. 2016‐51300‐25723/project accession no. 1010329), and by the Agriculture and Food Research Initiative Competitive Grant (grant no. 2022‐67015‐36073/project accession no. 1027785), from the USDA National Institute of Food and Agriculture.

Data Availability Statement

The data that support the findings of this study were available from the National Sheep Improvement Program. Restrictions apply to the wider availability of these data, as they were used with permission for the current study and are not publicly available. Data are however available from the authors upon reasonable request given the permission of the National Sheep Improvement Program.

References

  1. Al Kalaldeh, M. , Gibson J., Lee S. H., Gondro C., and van der Werf J. H. J.. 2019. “Detection of Genomic Regions Underlying Resistance to Gastrointestinal Parasites in Australian Sheep.” Genetics, Selection, Evolution 51: 37. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Alexander, G. , Stevens D., and Mottershead B.. 1983. “Problems in the Accurate Recording of Lambing Data.” Australian Journal of Experimental Agriculture 23: 361. [Google Scholar]
  3. Al‐Mamun, H. A. , Clark S. A., Kwan P., and Gondro C.. 2015. “Genome‐Wide Linkage Disequilibrium and Genetic Diversity in Five Populations of Australian Domestic Sheep.” Genetics, Selection, Evolution 47: 90. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Arisman, B. C. , Burke J. M., Morgan J. L. M., and Lewis R. M.. 2023. “Genotype by Environment Interaction and Heteroscedasticity Influence the Expression of Parasite Resistance in Katahdin Sheep.” Journal of Animal Science 101: skad228. 10.1093/jas/skad228. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Atlija, M. , Arranz J.‐J., Martinez‐Valladares M., and Gutiérrez‐Gil B.. 2016. “Detection and Replication of QTL Underlying Resistance to Gastrointestinal Nematodes in Adult Sheep Using the Ovine 50K SNP Array.” Genetics, Selection, Evolution 48: 4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Barbato, M. , Orozco‐terWengel P., Tapio M., and Bruford M. W.. 2015. “SNeP: A Tool to Estimate Trends in Recent Effective Population Size Trajectories Using Genome‐Wide SNP Data.” Frontiers in Genetics 6: 109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Becker, G. M. , Burke J. M., and Lewis R. M.. 2022. “706. Inbreeding and Effective Population Size of United States Katahdin Sheep.” In Proceedings of 12th World Congress on Genetics Applied to Livestock Production (WCGALP), edited by Veerkamp R. F. and de Haas Y., 3364. Wageningen, the Netherlands: Wageningen Academic Publishers. 10.3920/978-90-8686-940-4_706. [DOI] [Google Scholar]
  8. Becker, G. M. , Thorne J. W., Burke J. M., et al. 2024. “Genetic Diversity of United States Rambouillet, Katahdin and Dorper Sheep.” Genetics, Selection, Evolution 56: 1–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Brard, S. , and Ricard A.. 2015. “Is the Use of Formulae a Reliable Way to Predict the Accuracy of Genomic Selection?” Journal of Animal Breeding and Genetics 132: 207–217. [DOI] [PubMed] [Google Scholar]
  10. Burke, J. M. , and Miller J. E.. 2004. “Relative Resistance to Gastrointestinal Nematode Parasites in Dorper, Katahdin, and St. Croix Lambs Under Conditions Encountered in the Southeastern Region of the United States.” Small Ruminant Research 54: 43–51. [Google Scholar]
  11. Caballero, A. , Villanueva B., and Druet T.. 2021. “On the Estimation of Inbreeding Depression Using Different Measures of Inbreeding From Molecular Markers.” Evolutionary Applications 14: 416–428. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Chang, C. C. , Chow C. C., Tellier L. C., Vattikuti S., Purcell S. M., and Lee J. J.. 2015. “Second‐Generation PLINK: Rising to the Challenge of Larger and Richer Datasets.” GigaScience 4: 7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Dadousis, C. , Ablondi M., Cipolat‐Gotet C., et al. 2022. “Genomic Inbreeding Coefficients Using Imputed Genotypes: Assessing Different Estimators in Holstein‐Friesian Dairy Cows.” Journal of Dairy Science 105: 5926–5945. [DOI] [PubMed] [Google Scholar]
  14. Daetwyler, H. D. , Pong‐Wong R., Villanueva B., and Woolliams J. A.. 2010. “The Impact of Genetic Architecture on Genome‐Wide Evaluation Methods.” Genetics 185: 1021–1031. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Daetwyler, H. D. , Villanueva B., and Woolliams J. A.. 2008. “Accuracy of Predicting the Genetic Risk of Disease Using a Genome‐Wide Approach.” PLoS One 3: e3395. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Davenport, K. M. , Bickhart D. M., Worley K., et al. 2022. “An Improved Ovine Reference Genome Assembly to Facilitate In‐Depth Functional Annotation of the Sheep Genome.” GigaScience 11: 1–9. 10.1093/gigascience/giab096. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Davenport, K. M. , Hiemke C., McKay S. D., et al. 2020. “Genetic Structure and Admixture in Sheep From Terminal Breeds in the United States.” Animal Genetics 51: 284–291. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Davies, G. , Stear M. J., Benothman M., et al. 2006. “Quantitative Trait Loci Associated With Parasitic Infection in Scottish Blackface Sheep.” Heredity 96: 252–258. [DOI] [PubMed] [Google Scholar]
  19. Demars, J. , Cano M., Drouilhet L., et al. 2017. “Genome‐Wide Identification of the Mutation Underlying Fleece Variation and Discriminating Ancestral Hairy Species From Modern Woolly Sheep.” Molecular Biology and Evolution 34: 1722–1729. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Dementieva, N. V. , Mitrofanova O. V., Dysin A. P., et al. 2021. “Assessing the Effects of Rare Alleles and Linkage Disequilibrium on Estimates of Genetic Diversity in the Chicken Populations.” Animal 15: 100171. [DOI] [PubMed] [Google Scholar]
  21. Falconer, D. S. 1960. Introduction to Quantitative Genetics. Edinburgh; London, UK: Oliver And Boyd. [Google Scholar]
  22. FAO Animal Production and Health Division . 1998. Secondary Guidelines for Development of National Farm Animal Genetic Resources Management Plans: Management of Small Populations at Risk. Rome, Italy: Food and Agriculture Organization of the United Nations. [Google Scholar]
  23. Goddard, M. 2009. “Genomic Selection: Prediction of Accuracy and Maximisation of Long Term Response.” Genetica 136: 245–257. [DOI] [PubMed] [Google Scholar]
  24. Goddard, M. E. , Hayes B. J., and Meuwissen T. H. E.. 2011. “Using the Genomic Relationship Matrix to Predict the Accuracy of Genomic Selection.” Journal of Animal Breeding and Genetics 128: 409–421. [DOI] [PubMed] [Google Scholar]
  25. Gonyou, H. W. , and Stookey J. M.. 1985. “Behavior of Parturient Ewes in Group‐Lambing Pens With and Without Cubicles.” Applied Animal Behaviour Science 14: 163–171. [Google Scholar]
  26. Gutiérrez, J. P. , and Goyache F.. 2005. “A Note on ENDOG: A Computer Program for Analysing Pedigree Information.” Journal of Animal Breeding and Genetics 122: 172–176. [DOI] [PubMed] [Google Scholar]
  27. Hill, W. G. , and Robertson A.. 1968. “Linkage Disequilibrium in Finite Populations.” Theoretical and Applied Genetics 38: 226–231. [DOI] [PubMed] [Google Scholar]
  28. Hu, Z.‐L. , Fritz E. R., and Reecy J. M.. 2007. “AnimalQTLdb: A Livestock QTL Database Tool Set for Positional QTL Information Mining and Beyond.” Nucleic Acids Research 35: D604–D609. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Karaman, E. , Cheng H., Firat M. Z., Garrick D. J., and Fernando R. L.. 2016. “An Upper Bound for Accuracy of Prediction Using GBLUP.” PLoS One 11: e0161054. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Kennedy, B. W. , and Trus D.. 1993. “Considerations on Genetic Connectedness Between Management Units Under an Animal Model.” Journal of Animal Science 71: 2341–2352. [DOI] [PubMed] [Google Scholar]
  31. Kijas, J. W. , Porto‐Neto L., Dominik S., et al. 2014. “Linkage Disequilibrium Over Short Physical Distances Measured in Sheep Using a High‐Density SNP Chip.” Animal Genetics 45: 754–757. [DOI] [PubMed] [Google Scholar]
  32. Linck, E. , and Battey C. J.. 2019. “Minor Allele Frequency Thresholds Strongly Affect Population Structure Inference With Genomic Data Sets.” Molecular Ecology Resources 19: 639–647. [DOI] [PubMed] [Google Scholar]
  33. Lutaaya, E. , Misztal I., Bertrand J. K., and Mabry J. W.. 2001. “Inbreeding in Populations With Incomplete Pedigrees.” Journal of Animal Breeding and Genetics 116: 475–480. [Google Scholar]
  34. Malomane, D. K. , Reimer C., Weigend S., Weigend A., Sharifi A. R., and Simianer H.. 2018. “Efficiency of Different Strategies to Mitigate Ascertainment Bias When Using SNP Panels in Diversity Studies.” BMC Genomics 19: 22. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. McKenzie, F. C. , and Williams J.. 2015. “Sustainable Food Production: Constraints, Challenges and Choices by 2050.” Food Security 7: 221–233. [Google Scholar]
  36. McMillan, A. J. , Brown D. J., Burke J. M., Morgan J. L., and Lewis R. M.. 2022. “719. Cross‐Validation of Single‐Step Genetic Evaluation in U.S. Katahdin Sheep.” In Proceedings of 12th World Congress on Genetics Applied to Livestock Production (WCGALP), 2964–2967. Wageningen, Netherlands: Wageningen Academic Publishers. [Google Scholar]
  37. Meuwissen, T. H. 1997. “Maximizing the Response of Selection With a Predefined Rate of Inbreeding.” Journal of Animal Science 75: 934–940. [DOI] [PubMed] [Google Scholar]
  38. Meyermans, R. , Gorssen W., Buys N., and Janssens S.. 2020. “How to Study Runs of Homozygosity Using PLINK? A Guide for Analyzing Medium Density SNP Data in Livestock and Pet Species.” BMC Genomics 21: 94. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. National Research Council, Division on Earth and Life Studies, Board on Agriculture and Natural Resources, and Committee on the Economic Development and Current Status of the Sheep Industry in the United States . 2008. Changes in the Sheep Industry in the United States: Making the Transition From Tradition. Washington, D.C.: National Academies Press. [Google Scholar]
  40. Ngere, L. , Burke J. M., Morgan J. L. M., Miller J. E., and Notter D. R.. 2018. “Genetic Parameters for Fecal Egg Counts and Their Relationship With Body Weights in Katahdin Lambs.” Journal of Animal Science 96: 1590–1599. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Nilson, S. M. , Burke J. M., Murdoch B. M., Morgan J. L. M., and Lewis R. M.. 2024. “Pedigree Diversity and Implications for Genetic Selection of Katahdin Sheep.” Journal of Animal Breeding and Genetics 141: 304–316. 10.1111/jbg.12842. [DOI] [PubMed] [Google Scholar]
  42. Notter, D. R. 1998. “The U.S. National Sheep Improvement Program: Across‐Flock Genetic Evaluations and New Trait Development.” Journal of Animal Science 76: 2324–2330. [DOI] [PubMed] [Google Scholar]
  43. Notter, D. R. 2000. “Potential for Hair Sheep in the United States.” Journal of Animal Science 77: 1–8. [Google Scholar]
  44. Purcell, S. , Neale B., Todd‐Brown K., et al. 2007. “PLINK: A Tool Set for Whole‐Genome Association and Population‐Based Linkage Analyses.” American Journal of Human Genetics 81: 559–575. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Rafter, P. , McHugh N., Pabiou T., and Berry D. P.. 2022. “Inbreeding Trends and Genetic Diversity in Purebred Sheep Populations.” Animal 16: 100604. [DOI] [PubMed] [Google Scholar]
  46. Raj, A. , Stephens M., and Pritchard J. K.. 2014. “fastSTRUCTURE: Variational Inference of Population Structure in Large SNP Data Sets.” Genetics 197: 573–589. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Reid, J. M. , Keller L. F., Marr A. B., Nietlisbach P., Sardell R. J., and Arcese P.. 2014. “Pedigree Error Due to Extra‐Pair Reproduction Substantially Biases Estimates of Inbreeding Depression.” Evolution 68: 802–815. [DOI] [PubMed] [Google Scholar]
  48. Silva, M. V. B. , Sonstegard T. S., Hanotte O., et al. 2012. “Identification of Quantitative Trait Loci Affecting Resistance to Gastrointestinal Parasites in a Double Backcross Population of Red Maasai and Dorper Sheep.” Animal Genetics 43: 63–71. [DOI] [PubMed] [Google Scholar]
  49. Simianer, H. , Büttgen L., Ganesan A., Ha N. T., and Pook T.. 2021. “A Unifying Concept of Animal Breeding Programmes.” Journal of Animal Breeding and Genetics 138: 137–150. [DOI] [PubMed] [Google Scholar]
  50. Taberlet, P. , Coissac E., Pansu J., and Pompanon F.. 2011. “Conservation Genetics of Cattle, Sheep, and Goats.” Comptes Rendus Biologies 334: 247–254. [DOI] [PubMed] [Google Scholar]
  51. Thorne, J. W. , Murdoch B. M., Freking B. A., et al. 2021. “Evolution of the Sheep Industry and Genetic Research in the United States: Opportunities for Convergence in the Twenty‐First Century.” Animal Genetics 52: 395–408. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. USDA National Agriculture Statistics Service . 2020. “Sheep and Goats.” https://downloads.usda.library.cornell.edu/usda‐esmis/files/000000018/n296xf83n/m900pb410/shep0120.pdf.
  53. van der Werf, J. 2013. “Genomic Selection in Animal Breeding Programs.” In Genome‐Wide Association Studies and Genomic Prediction, edited by Gondro C., van der Werf J., and Hayes B., 543–561. Totowa, NJ: Humana Press. [DOI] [PubMed] [Google Scholar]
  54. Villanueva, B. , Fernández A., Saura M., et al. 2021. “The Value of Genomic Relationship Matrices to Estimate Levels of Inbreeding.” Genetics, Selection, Evolution 53: 42. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Vostry, L. , Milerski M., Schmidova J., and Vostra‐Vydrova H.. 2018. “Genetic Diversity and Effect of Inbreeding on Litter Size of the Romanov Sheep.” Small Ruminant Research 168: 25–31. [Google Scholar]
  56. Wang, J. 2016. “Pedigrees or Markers: Which Are Better in Estimating Relatedness and Inbreeding Coefficient?” Theoretical Population Biology 107: 4–13. [DOI] [PubMed] [Google Scholar]
  57. Wiggans, G. R. , Sonstegard T. S., VanRaden P. M., et al. 2009. “Selection of Single‐Nucleotide Polymorphisms and Quality of Genotypes Used in Genomic Evaluation of Dairy Cattle in the United States and Canada.” Journal of Dairy Science 92: 3431–3436. [DOI] [PubMed] [Google Scholar]
  58. Wildeus, S. 1997. “Hair Sheep Genetic Resources and Their Contribution to Diversified Small Ruminant Production in the United States.” Journal of Animal Science 75: 630–640. [DOI] [PubMed] [Google Scholar]
  59. Wilson, C. S. , Petersen J. L., Blackburn H. D., and Lewis R. M.. 2022. “Assessing Population Structure and Genetic Diversity in US Suffolk Sheep to Define a Framework for Genomic Selection.” Journal of Heredity 113: 431–443. [DOI] [PubMed] [Google Scholar]
  60. Wilson, D. E. , and Morrical D. G.. 1991. “The National Sheep Improvement Program: A Review.” Journal of Animal Science 69: 3872–3881. [DOI] [PubMed] [Google Scholar]
  61. Wray, N. R. , and Goddard M. E.. 1994. “Increasing Long‐Term Response to Selection.” Genetics, Selection, Evolution 26: 1–21. [Google Scholar]
  62. Wright, S. 1922. “Coefficients of Inbreeding and Relationship.” American Naturalist 56: 330–338. [Google Scholar]
  63. Yu, H. , Spangler M. L., Lewis R. M., and Morota G.. 2017. “Genomic Relatedness Strengthens Genetic Connectedness Across Management Units.” G3 7: 3543–3556. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Figure S1.

JBG-142-408-s001.zip (521.2KB, zip)

Table S1.

JBG-142-408-s002.docx (35.8KB, docx)

Data Availability Statement

The data that support the findings of this study were available from the National Sheep Improvement Program. Restrictions apply to the wider availability of these data, as they were used with permission for the current study and are not publicly available. Data are however available from the authors upon reasonable request given the permission of the National Sheep Improvement Program.

Articles from Journal of Animal Breeding and Genetics are provided here courtesy of Wiley

RESOURCES