Introduction. In modern animal husbandry, the exterior of cattle is an important production indicator directly affecting health, productivity, and reproductive function [1]. Body constitution indices are taken into account in dairy cow breeding among the primal indicators to assess the non-productive traits [2]. Gibson et al. believe that height of a calf is related to the lactation efficiency of an adult cow: taller calves will produce more milk in the future through several lactations compared to the shorter calves, live weight considered [3]. The results of the study conducted by Schmidtmann et al. show that larger animals with visibly protruding bones and a low BCS (Body Condition Score) are more susceptible to metabolic disorders [4].

Exterior traits can be decisive for the development of pathological conditions in animals. For example, sacrum anatomy features can be related to the risk of dystocia (pathological or difficult parturition) [5]. Significantly higher incidence of dystocia has been found in cows with the sacrum size both narrower and, on the contrary, wider than optimal for calving [6]. The influence of a breed on the incidence of dystocia is explained by the differences in a relative body weight at birth, pelvis anatomy, and significant variations in dimensions of pelvis in some breeds [5]. Due to modern breeding focusing on selection of the herd specimens with the highest milk yield and long-term economic productivity, dystocia remains a serious problem with incidence rate estimated at 10 to 50% [7].

The aim of the study is to identify associations of single-nucleotide polymorphisms with pelvic exterior parameters in Holstein cows that are important prerequisites for forming susceptibility to easy or difficult calving. The data obtained will make it possible to determine the genetic markers to be included into the herd improvement breeding programs.

Materials and Methods. The study was carried out at the Department of Genomic Research and Animal Breeding of Ural Research Veterinary Institute, structural division of Ural Federal Agrarian Research Center of the Ural Branch of the Russian Academy of Sciences, in the period from 2023 to 2025. PCR tests of DNA samples from Holstein cows (n=155) from two cattle breeding agricultural enterprises (AEs) (AE No. 1, n=70; AE No. 2, n=85) were performed. The research sample included animals with one or more lactations. The research was held in compliance with the ethical standards, the studies (e.g., blood sampling from the caudal veins of animals) were approved by the Bioethics Committee of Ural Federal Agrarian Research Center, Ural Branch of the Russian Academy of Sciences (Protocol No. 136 of October 13, 2025).

Exterior traits of the animals were assessed across three dimensions: sacrum height, hip width, and ischial tuberosity width. Statistical data for a phenotype —mean values and standard deviations—were calculated using STATISTICA 12 software (StatSoft Inc, USA).

13 Single Nucleotide Polymorphisms (SNPs) were genotyped across entire sample of cows. To perform association tests, the following polymorphic markers were studied: rs134055603, rs137396952, rs109452259, rs110347054, rs110352004, rs133674837, rs109529386, rs109279094, rs109611915, rs134668940, rs41567027, rs43038601, rs41255693, which had previously shown significant associations with a number of physiologically valuable parameters in cattle during GWAS studies:

– productive longevity [8];

– easy calving [9][10];

– body weight [11];

– carcass length after cutting [12];

– and other parameters of productivity [10][13].

The article presents both authors’ own findings based on TaqMan genotyping technology and earlier described test systems (Table 1).

Based on the genetic data obtained, the relationship of each polymorphic locus with a phenotype was assessed in separate. Five patterns of trait inheritance were assessed using the SNPStats web tool (Institut Català d'Oncologia, Spain) [19] and the SNPassoc package [20] in R programming environment (Autonomous University of Barcelona, Spain): codominant, dominant, recessive, overdominant, and log-additive; the effect size was interpreted as the odds ratio (the differences in the values of the exterior traits) with a 95% confidence interval compared to alternative allele homozygote.

To conduct association studies using the SNPassoc package in the R programming environment, the WGassociation function was used. To test for the presence of “gene-gene” interactions, the interactionPval (Phen, data=data, model=’co’) function was used. Association tests were performed across three pelvic dimensions: sacrum height, hip width, and ischial tuberosity width; the presence of “gene-gene” interactions was also assessed for each of the exterior traits. For a more detailed study of allele combinations contributing to the formation of a trait, taking into account “gene-gene” interactions, a linear model was created: lm(“Phen ~”, “SNP1”, “*”, “SNP2”, data=data). With regard to epistatic interactions, this model envisaged sorting through each combination of genotypes and determining mean value of each of these combinations for a phenotype. A post-hoc Bonferroni correction was performed for each analysis

Results. The mean values and standard deviations for sacrum height in Holstein cows were determined to be 161.7±5.3 cm; for hip width — 41.1±3.9 cm; and for ischial tuberosity width — 22.4±2.1 cm.

Larger values of pelvic width proved to ensure sufficient space for successful development of a fetus and easy calving, and necessary space between hindlimbs to ensure a more voluminous cup-shaped udder [21]. Thus, higher values of each of the three exterior traits are considered a favourable phenotype.

Significant effects of rs109452259 SNP on hip width in Holstein cows were acknowledged in codominant, dominant, and log-additive inheritance patterns. The lowest AIC value for rs109452259 was found in the dominant inheritance pattern, with the C/A and A/A genotypes having 2.17 cm difference from C/C; p-value=0.0006.

When studying a phenotype inheritance mechanism, an interesting feature was revealed, i.e. the presence of “gene-gene” interactions among polymorphisms. In the present study, the epistatic interactions for the pairs of rs134055603–rs43038601 and rs134055603–rs137396952 polymorphisms were identified according to the codominant inheritance pattern (Fig. 1). The significance level for these pairs was corrected to α<0.00032.

A statistically significant epistatic interaction was found for rs134055603 polymorphism. This SNP has a significantly positive contribution to sacrum height only in the presence of A* allele in a genotype identified by rs43038601 polymorphism, while its absence (the combination of G/G and C/C genotypes identified by rs134055603 and rs43038601, respectively) has a negative effect on the phenotype (Fig. 2). It is worth noting that the number of specimens of A/A genotype identified by rs43038601 polymorphism for each of the three combinations was ≤ 3. Therefore, interpretation of these genotype combinations is impossible. For each of the other six combinations, the sample size was ≥ 10.

Epistatic interactions of opposite nature for 134055603 with rs137396952 polymorphism were found. The combination of A* allele in genotype of rs134055603 and the C* allele in genotype of rs137396952 had a positive effect on sacrum height (Fig. 3). It is worth emphasising that the number of specimens with A/A genotype identified by rs43038601 polymorphism for each of the three combinations was ≤ 3.

Discussion and Conclusion. During the study conducted in 155 Holstein cows, the mean values and standard deviations for sacrum height (161.7±5.3 cm), hip width (41.1±3.9 cm), and ischial tuberosity width (22.4±2.1 cm) were determined. Genotyping of 13 polymorphisms was conducted for each specimen. It was revealed that in Holstein cows, genotyping of rs109452259 requires negative selection by C* allele to increase the hip width. Epistatic interaction of rs134055603 and rs43038601, probably of synergistic nature for G/G and C/A genotypes respectively, was identified. Epistatic interaction of rs134055603 and rs137396952, probably of synergistic nature for the C/A and C/C genotypes respectively, was identified. The data obtained during the study are significant for the observed-level statistical interaction; for confirmation of the biological mechanisms, further functional studies are required.

The potential limitations of the study include possible influence of the unaccounted confounding factors, such as, variability of the studied groups of animals by age, live body weight, lactation number, and origin (of an agricultural enterprise). The above mentioned limitations could have influenced the results obtained and their subsequent interpretation.