The effects of genotype-by-environment interactions on body condition score across three winter supplemental feed environments in a composite beef cattle breed in Montana

Abstract

Cattle operations in the Northern Great Plains region of the United States face extreme cold weather conditions and require nutritional supplementation over the winter season in order for animals to maintain body condition. In cow–calf operations, body condition scores (BCS) measured at calving and breeding have been shown to be associated with several economically important health and fertility traits, so maintenance of BCS is both an animal welfare and economic concern. A low-to-medium heritability has been found for BCS when measured across various production stages, indicating a large environmental influence but sufficient genetic basis for selection. The present study evaluated BCS measured prior to calving (late winter) and breeding (early summer) under three winter supplementation environments in a multitrait linear mixed model. Traits were discretized by winter supplementation and genetic correlations between environments were considered a reflection of evidence for genotype-by-environment interactions between BCS and diet. Winter supplementation treatments were fed October through April and varied by range access and protein content: 1) feedlot environment with approximately 15% crude protein (CP) corn/silage diet, 2) native rangeland access with 1.8 kg of an 18% CP pellet supplement, and 3) native rangeland access with a self-fed 50% CP and mineral supplement. A total of 2,988 and 2,353 records were collected across multiple parities on 1,010 and 800 individuals for prebreeding and precalving BCS, respectively. Heifers and cows came from a composite beef cattle breed developed and maintained by the USDA Fort Keogh Livestock and Range Research Laboratory near Miles City, Montana. Genetic correlations between treatments 1 and 2, 1 and 3, and 2 and 3 were 0.98, 0.78, and 0.65 and 1.00, 0.98, and 0.99 for precalving and prebreeding BCS, respectively. This provides moderate evidence of genotype-by-environment interactions for precalving BCS under treatment 3 relative to treatments 1 and 2, but no evidence for genotype-by-environment interactions for prebreeding BCS. Treatment 3 differed substantially in CP content relative to treatments 1 and 2, indicating that some animals differ in their ability to maintain BCS up to spring calving across a protein gradient. These results indicate the potential for selection of animals with increased resilience under cold weather conditions and high protein, restricted energy diets to maintain BCS.

This study presents evidence of the effects of G × E interactions on body condition scores measured at calving across a winter supplementation protein gradient in a composite breed of beef cattle in Montana. This genetic variation may allow producers the opportunity to select animals more resilient to winter conditions using fewer supplementary feed resources.

INTRODUCTION

Calf–cow operations in the Northern Great Plains region of the United States face sustained temperatures close to or below freezing throughout the winter months. A general management practice in this region is to supplement the forage available to cattle in feedlots or on native range to maintain body condition and avoid losses in productivity that may lead to increased culling. Cold weather conditions can induce physiological adaptations at the metabolic level that increase the baseline metabolic rate to maintain core body temperatures which can stimulate the animal’s appetite to meet increased energy demands (Young, 1983). These adaptations may impact nutrient utilization as well; Mader et al. (2001) found that cattle exposed to extreme cold stress benefited more as the dietary availability of starch increased relative to the roughage amount. To optimize herd reproductive performance in spring and summer, it has been recommended that cows in poor body condition entering winter be supplemented with a high-energy, low-protein (15% to 30% CP) diet (Herd and Sprott, 1986).

Body condition scores (BCS) measured at calving and breeding have been linked with several economically important traits in calf–cow production systems, including pregnancy rate, interval from calving to first estrus, calving interval, and calf vitality (Herd and Sprott, 1986; Selk et al., 1988). Pregnancy rates have been shown to increase with higher BCS (Herd and Sprott, 1986), with a trend approximately following a sigmoid curve for precalving BCS between 3 and 7 according to Selk et al. (1988). Rae et al. (1993) found a statistically significant association between body condition and pregnancy rate and showed that cows with BCS ≤ 4 at pregnancy check had a 59% conception rate compared to a 90% conception rate for cows with BCS > 4. Ciccioli et al. (2003) fed cows one of two diets differing calorically postcalving and found that cows fed the calorically deficient diet had significantly lower conception rates and higher days from calving to first estrus. Statistically significant differences in the milk yield between high, medium, and low BCS classes in the Spanish Parda de Montaña beef cattle breed were described by Cortés-Lacruz et al. (2017), potentially impacting the maternal environment and by extension calf growth and performance. High BCS have also been found to have a statistically significant association with increased calving difficulty in beef heifers (Paputungan et al., 1994), which may factor into culling decisions and by extension longevity of cows.

Low-to-moderate heritabilities for BCS that vary depending on the stage of production, age of the animal, and consistency in scoring of handlers have been reported for both beef and dairy cattle (Veerkamp et al., 2001; Nephawe et al., 2004; Weik et al., 2021; Ribeiro et al., 2022). This genetic variation may be suitable for selection of individuals with a better genetic basis for maintaining sufficiently high BCS (generally considered to be a moderate score comparable to 5 or 6 on a 1 to 9 scale) at critical production stages in a calf–cow system, though it does not distinguish to what degree individuals maintaining higher BCS score utilize their feed more efficiently or increase total feed consumed.

A composite breed of beef cattle (Composite Gene Combination, CGC) developed and maintained by the USDA-ARS Livestock and Range Research Laboratory (Newman et al., 1993a, 1993b) near Miles City, Montana, has been the subject of winter supplementation feed trials that vary in forage and protein composition to evaluate the impact on performance in a cow–calf operation. Previous studies conducted for this population have reported evidence of genotype-by-environment (G × E) interactions across nutritional environments for several traits, including average daily gain (Hay and Roberts, 2018) and calf growth (Hay and Roberts, 2019).

The present study aimed to investigate the potential existence of G × E interactions between three winter dietary supplemental treatments and BCS collected prior to calving and breeding in the CGC composite breed. Evidence for G × E interactions was evaluated via genetic correlations between environments in a multitrait model.

MATERIALS AND METHODS

Data were collected on heifers and cows from the CGC breed at the USDA, ARS, Fort Keogh Livestock and Range Research Laboratory near Miles City, Montana. CGC is a composite breed of beef cattle initially developed by crossing Red Angus dams with Charolais and Tarentaise sires with the goal of achieving an approximate ½ Red Angus, ¼ Charolais, and ¼ Tarentaise breed composition, full details of which can be found in Newman et al. (1993a, 1993b). A total of 1,026 heifers and cows born between 2012 and 2019 were evaluated for BCS on a 1 to 9 scale (Herd and Sprott, 1986) prior to breeding in late May to early June and prior to calving in late February to early March. Each BCS record was collected once per parity throughout the duration of the animal’s herd life, with records collection occurring between 2014 and 2021. A total of 2,988 and 2,353 records and average of 2.96 and 2.94 records per cow (Table 1) were collected for prebreeding and precalving BCS, respectively. For the majority of the study, the same person conducted the body condition scoring for the cattle.

Table 1.

Summary statistics for prebreeding and precalving BCS data, pedigree, and genotypes

	Prebreeding BCS	Precalving BCS
Records, n
Total	2,988	2,353
Treatment 1	1,003	807
Treatment 2	968	752
Treatment 3	1,017	794
Min/cow	1	1
Max/cow	8	8
Average/cow	2.96	2.94
Animals with records, n
Heifers and cows	1,010	800
Treatment 1	324	266
Treatment 2	342	274
Treatment 3	344	260
Daughters per sire
Min	1	1
Max	35	26
Average	9.2	7.3
Pedigree, n
Total animals	5,822	5,813
Sires
Total	202	202
Have daughter(s) with record(s)	108	108
Dams
Total	1,740	1,737
Have daughter(s) with record(s)	618	536
Genotypes (after QC)
SNPs	40,533
Total animals	5,468
Sires	106	106
Dams	588	508
Heifers and cows with record(s)	988	787

Beginning in their first winter after weaning and continuing throughout their lifetime in the herd, heifers and cows were fed one of three dietary treatments from the months of October through April. These treatments are described as follows: 1) confinement to outdoor lots (drylots) with ad libitum access to a corn silage/ground alfalfa/mineral supplemental diet that provides 1.4 Mcal/kg NEm and 0.8 Mcal/kg NEg (DM basis) and on average approximately 15% crude protein (CP) content; 2) access to dormant native range along with 1.8 kg per head daily of a ground alfalfa/energy grains pellet supplement that provides 1.5 Mcal/kg NEm and 1.0 Mcal/kg NEg (as fed basis) and on average approximately 18% CP; and 3) access to dormant native range plus ad libitum access to a high protein (minimum 61% CP) mineral mix (Mulliniks et al., 2012) provided using GrowSafe feed intake systems. Treatments differed primarily in whether energy was sourced from supplemented ground alfalfa/silage/energy grains or native rangeland, with individuals in treatment 3 provided a concentrated protein supplement to increase forage intake (Kartchner, 1980). All individuals were managed on the same winter supplementation treatment throughout their productive herd life.

To accommodate the availability of resources, the composition of treatment 1 varied across years between 63% to 81% corn silage and 17% to 37% ground alfalfa, with the remaining consisting of a protein/mineral mix or commercial mineral mix (VitaFerm Cattleman’s Blend) to achieve target nutritional composition. The protein/mineral mix was composed of (DM basis) 9% barley, 4.2% soybean meal, 0.9% urea, 0.5% calcium carbonate, 0.2% NaCl, 0.1% vitamin mix (44,000,000 IU/kg vitamin A, 880,000 IU/kg vitamin D, and 880 IU/kg vitamin E), and 0.1% trace mineral (20.0% Mg, 0.2% K, 2.6% S, 18,000 ppm Cu, 60,000 ppm Zn, 40,000 ppm Fe, 300 ppm Se, 60,000 ppm Mn, 180 ppm Co, and 1,140 ppm I). This mineral mix was also the protein/mineral supplementation administered to animals in treatment 3. The diet for treatment 2 was composed of 60% alfalfa and 40% energy grains, with the exception of Fall 2015 through January 2017 when resources necessitated a formulation consisting of 61% wheat midds, 20% malt sprouts, 5% alfalfa, 5% molasses, 4.7% dried-distillers grain, and 4.3% mineral–vitamin mix.

A total of 5,468 individuals were genotyped using the BovineSNP50 chip, with 40,533 SNPs retained following quality control measures for low minor allele frequency (<0.05) and high rate of missingness (>0.1). Genotypes were also evaluated for deviations from Hardy–Weinberg equilibrium and parent-progeny Mendelian conflicts, but no SNPs or animals were filtered based on these criteria. Complete summary statistics for genotype data can be found in Table 1.

For the purposes of evaluating G × E interactions between nutritional environment and BCS, a three-trait model was assumed,

$$\left[\begin{array}{c}{y}_1\\ y_2\\ y_3\end{array}\right]=\left[\begin{array}{ccc}{X}_1& 0& 0\\ 0& X_2& 1\\ 0& 0& X_3\end{array}\right]\left[\begin{array}{c}{b}_1\\ b_2\\ b_3\end{array}\right]+\left[\begin{array}{ccc}{Z}_1& 0& 0\\ 0& Z_2& 0\\ 0& 0& Z_3\end{array}\right]\left[\begin{array}{c}{u}_1\\ u_2\\ u_3\end{array}\right]+\left[\begin{array}{ccc}{W}_1& 0& 0\\ 0& W_2& 0\\ 0& 0& W_3\end{array}\right]\left[\begin{array}{c}{p}_1\\ p_2\\ p_3\end{array}\right]+\left[\begin{array}{c}{e}_1\\ e_2\\ e_3\end{array}\right],$$ (1)

where y_i corresponds to records observed for the ith winter supplementation treatment; X_i, Z_i, and W_i represent design matrices for fixed, random additive genetic, and random permanent environment effects, respectively; b_i, u_i, and p_i are solutions for fixed, random additive genetic, and random permanent environment effects, respectively; and e_i are residuals. Fixed effects for both prebreeding and precalving BCS included mean, birth year, and parity. Additive genetic relationships were modeled by pedigree relationships for variance component estimation and genomic relationships for prediction of fixed and random effects,

$$u\sim N(0,G_0\otimes G),$$

where G was either the numerator relationship matrix or genomic relationship matrix (VanRaden, 2008) and G₀ is a matrix of additive genetic (co)variances between traits,

$$G_0=\left[\begin{array}{ccc}{\sigma }_{u1}^2& {\sigma }_{u12}& {\sigma }_{u13}\\ {\sigma }_{u12}& {\sigma }_{u2}^2& {\sigma }_{u23}\\ {\sigma }_{u13}& {\sigma }_{u23}& {\sigma }_{u3}^2\end{array}\right],$$

Permanent environment effects were assumed to follow,

$$p\sim N(0,P_0\otimes I),$$

and residuals,

$$e\sim N(0,R_0\otimes I),$$

$$where{P}_0=\left[\begin{array}{ccc}{\sigma }_{p1}^2& {\sigma }_{p12}& {\sigma }_{p13}\\ {\sigma }_{p12}& {\sigma }_{p2}^2& {\sigma }_{p23}\\ {\sigma }_{p13}& {\sigma }_{p23}& {\sigma }_{p3}^2\end{array}\right]and{R}_0=\left[\begin{array}{ccc}{\sigma }_{e1}^2& 0& 0\\ 0& {\sigma }_{e2}^2& 0\\ 0& 0& {\sigma }_{e3}^2\end{array}\right].$$

Variance component estimation and prediction of fixed and random effects were conducted in REMLF90 and BLUPF90 (Misztal et al., 2002) with a single-step GBLUP model (Aguilar et al., 2010; Christensen and Lund, 2010), respectively.

Evidence for G × E interactions was evaluated based on estimated genetic correlations between traits grouped by winter supplementation treatments in this model, with a greater departure from a genetic correlation of 1 representing increasing evidence of G × E interactions across nutritional environments.

RESULTS

The distributions of prebreeding and precalving BCS across winter supplementation treatments are shown in Fig. 1a and b, respectively. The distribution of BCS is fairly similar for both treatments with the highest incidence occurring for a body condition of 5, generally considered to be the ideal for body condition. There was a lower incidence of body conditions above 5 observed for precalving relative to prebreeding BCS. The mean (standard deviation) of precalving BCS were 4.79 (0.42), 4.86 (0.40), and 4.82 (0.42) for treatments 1, 2, and 3, respectively; these were lower than the mean (standard deviation) of prebreeding BCS of 5.01 (0.68), 4.94 (0.63), and 4.85 (0.61) for treatments 1, 2, and 3, respectively.

Figure 1.

Distribution of body condition scores measured prior to breeding (a) and calving (b) grouped by treatment.

Estimates of univariate variance components, heritabilities, and repeatabilities for prebreeding and precalving BCS are presented in Table 2. Heritabilities for both prebreeding and precalving BCS were low-to-moderate at 0.18 and 0.20, respectively, consistent with estimates published previously in other beef populations (Nephawe et al., 2004; Weik et al., 2021; Ribeiro et al., 2022). The repeatability for precalving BCS (0.26) was found to be higher than that for prebreeding BCS (0.18) due to an estimated permanent environment variance near zero for the latter. When prebreeding and precalving BCS were evaluated jointly in a two-trait model, the estimated additive genetic correlation was near unity at 0.99, indicating a similar additive genetic mechanism for both traits.

Table 2.

Variance components estimated in single-trait model

	Genetic variance	Permanent environment variance	Residual variance	Heritability	Repeatability
Prebreeding BCS	0.040 (5.20 × 10−2)	0.63 × 10−4 (0.17 × 10−3)	0.18 (5.39 × 10−2)	0.18	0.18
Precalving BCS	0.030 (8.05 × 10−2)	0.011 (6.53 × 10−2)	0.12 (4.00 × 10−2)	0.20	0.26

Table 3 shows estimates of genetic covariances and correlations, heritability, and repeatability for the multitrait model intended to measure G × E interactions. Heritabilities for the three treatments under prebreeding BCS ranged from 0.14 to 0.18, with treatment 1 having the lowest estimated heritability. Precalving BCS heritabilities ranged from 0.17 to 0.28 with treatments 2 and 1 having the lowest and highest heritabilities, respectively. The heritability of treatment 2 was noticeably lower than those for treatments 1 and 3, though this may be potentially explained as an artifact of the relatively small sample size (Table 1). Repeatabilities across treatments for precalving BCS were fairly similar, ranging from 0.25 to 0.29.

Table 3.

Heritabilities (diagonal), additive genetic covariances (below diagonal), genetic correlations (above diagonal), and repeatabilities (bottom row) for prebreeding and precalving BCS in a three-trait model

	Prebreeding BCS			Precalving BCS
	Treatment 1	Treatment 2	Treatment 3	Treatment 1	Treatment 2	Treatment 3
Treatment 1	0.14	1.00	0.98	0.28	0.98	0.78
Treatment 2	0.034 (8.96 × 10−2)	0.18	0.99	0.033 (7.47 × 10−2)	0.17	0.65
Treatment 3	0.033 (7.56 × 10−2)	0.039 (7.49 × 10−2)	0.18	0.032 (9.65 × 10−2)	0.019 (8.31 × 10−2)	0.23
Repeatability	0.17	0.19	0.18	0.29	0.25	0.25

Genetic correlations between treatments under prebreeding BCS were near unity, with the lowest genetic correlation being 0.98 between treatments 1 and 3. This indicates little to no reranking of genomic estimated breeding values (GEBVs) across treatments and no evidence for the existence of G × E between prebreeding BCS and the dietary environments evaluated in this study. In contrast, genetic correlations for precalving BCS between treatments 1 and 3 and 2 and 3 were 0.78 and 0.65, respectively. A threshold of 0.8 for genetic correlation between environments is commonly used as a benchmark in the literature to indicate the presence of G × E interactions, and so the genetic correlations reported here indicate a modest reranking of animals when comparing treatment 3 to the other 2 diets. With a genetic correlation of 0.98 between treatments 1 and 2, however, there was no evidence for reranking of animals under these dietary treatments.

Because no cows have records for more than one treatment, GEBVs of sires with daughters across multiple treatments are expected to be more reliable. The distribution of the number of daughters observed in each treatment across sires is shown in Fig. 2. Several sires have at least 10 daughters with a fairly even spread across treatments and the ranking of GEBVs of these sires across treatments is shown in Fig. 3a and b for prebreeding and precalving BCS, respectively. Consistent with the estimated genetic correlations between treatments (≥0.98), little reranking across treatments is observed for prebreeding BCS. Among the 24 sires represented in Fig. 3a, the Spearman rank correlations between GEBVs for prebreeding BCS are 1.00, 0.999, and 0.999 for treatments 1 and 2, 1 and 3, and 2 and 3, respectively.

Figure 2.

Number of daughters in each treatment per sire.

Figure 3.

EBVs of sires with at least 10 daughters across treatments for body condition score measured prior to breeding (a) and calving (b). Sires with at least 5 daughters in each treatment are shown by a dashed line.

In contrast, while precalving BCS GEBVs are fairly consistent across treatments for some sires, Fig. 3b shows significant reranking for others when comparing solutions under treatments 1 and 2 to treatment 3. The corresponding Spearman rank correlations are 0.976, 0.806, and 0.709 between treatments 1 and 2, 1 and 3, and 2 and 3, respectively, indicating large rank changes for some sires under treatment 3 relative to treatments 1 and 2.

DISCUSSION

The reranking of GEBVs for precalving BCS under treatment 3 relative to treatments 1 and 2 indicates some variation in the genetic basis contributing to performance under this environment compared to the others. Treatment 3 differs from treatments 1 and 2 in that little supplemental energy was provided and individuals in this treatment relied on forage from native rangeland to meet energy requirements. While native rangeland may be suitable to meet energy requirements of cattle over winter, insufficient protein content of forage can result in reduced forage intake (Kartchner, 1980). A concentrated protein supplement (minimum 61% CP) was provided ad libitum to encourage increased forage intake and, though individual intake of this supplement was not evaluated in the present study, variation in intake of this supplement may potentially explain differences in performance under treatment 3. Alternatively, it is plausible that genetic variation exists in the degree to which forage intake increases with increased protein intake. The high genetic correlations across winter supplementation treatments for prebreeding BCS can likely be explained by a sufficient amount of time having passed between cessation of supplementation provision and measurement of prebreeding BCS for cows to have recovered differences in BCS attributable to the treatments. While the average precalving BCS (4.82) across all treatment groups is lower than that of prebreeding BCS (4.93), this is a relatively small difference of approximately 3.74 kg considering 1 body condition score is equivalent to 34 kg on a BCS 1 to 9 scale (Kunkle et al., 1994). Precalving BCS measurements were collected in February or March before the dietary trials had ended in April, so animals would still be subject to effects on body condition induced by dietary differences. Lake et al. (2005) found a statistically significant higher pregnancy rate (P < 0.02) in 3-year-old Angus × Gelbvieh cows with a BCS of 6 at parturition relative to those with a BCS of 4, further supporting the importance of BCS at calving on future reproductive performance. Future studies may investigate the relationship between precalving BCS across treatments with pregnancy rate and other measures of reproductive performance.

Previous studies have reported statistically significant differences in pregnancy rates and other fertility traits between groups that differed in BCS measured at calving (Rae et al., 1993; Ciccioli et al., 2003). By extension, while G × E interactions found for precalving BCS were not observed for prebreeding BCS, this should not necessarily be extrapolated to an absence of differences in performance for traits measured postcalving. For the same population, Hay and Roberts (2019) found evidence of G × E interactions in the maternal genetic environment for birth, weaning, and yearling weights across winter supplementation environments that varied in protein percentage. It may be useful in future studies to investigate genetic correlations of body condition with fertility and performance traits across a protein gradient environment to determine how individuals that demonstrate a more robust genetic basis for maintenance of body condition under restricted nutrition perform with regards to correlated traits.

For spring-calving cows, the general advice for weathering heifers and cows over winter is to provide sufficient winter supplementation such that BCS of 5 or more can be maintained into spring with the purpose of maximizing pregnancy rates, maintaining a 12-month calving interval, and optimizing the maternal environment that can be provided. While this approach benefits short-term herd performance, it minimizes variation attributable to robustness under environmental challenges and by extension reduces the opportunity to select for more resilient individuals. The present study has shown the existence of genetic variation for both precalving and prebreeding BCS and evidence for differences in the genetic basis of precalving BCS across environments differing in whether animals receive high-energy supplemental feed or rely on forage from native rangeland to meet energy needs over winter. It may then be feasible for producers to customize winter management based on individual performance under different environments or manage all animals under the conditions best suited for their practice and select for animals best suited to those management practices.

The present analysis was conducted on a composite breed in development since 1978 (Newman et al., 1993a). Composite breeds have been suggested as a preferable alternative to terminal and rotational crossbreeding systems for the purposes of combining beneficial additive characteristics from complementary breeds while harnessing the benefits of heterosis in small cattle production systems (Gregory and Cundiff, 1980). Though not investigated in the current study, it is plausible that differences in performance across nutritional environments could be attributable to variation in breed contributions, including heterotic effects from combinations of alleles originating in different breeds. Gregory et al. (1994a, 1994b) found in a comparison of composites to founder breeds, variation in several traits was attributable to heterosis even several generations into composite formation. Hay et al. (2022) investigated the breed composition across individuals in the CGC breed; while the variation of breed contribution was low among modern generations, the observed variation may be sufficient to explain reranking of sires across environments. Further studies may consider whether haplotypes originating in a particular founding breed and heterosis from combinations of alleles from different breeds are associated with differences in performance across nutritional environments.

CONCLUSIONS

The current study presented modest evidence of G × E interactions, based on a genetic correlation threshold of 0.8, across winter supplementation treatments differing in protein content in a composite breed of beef cattle. Devising an appropriate composition of winter supplementation that balances nutritional needs of animals against cost to producers is an important consideration in cow–calf production systems given the known impact of BCS at critical production stages on health and fertility traits. The results presented here suggest that some individuals may have a genetic basis for increased resilience in maintaining body condition under nutritional challenge (high protein/low TDN) in cold weather conditions. Several sires showed significant reranking between treatments that differed in protein content and supplemental energy provided under drylot or rangeland environments, indicating that some individuals may have a genetic basis better suited for wintering in a rangeland environment with limited supplemental resources. By extension, there may be potential for genetic selection of animals that are better able to maintain body condition leading into the breeding season under typical rangeland conditions in Montana over winter while necessitating only protein and mineral supplementation. This presents a potential avenue for producers to develop a herd better able to utilize readily available rangeland resources with lower supplemental costs.