Gestation length in dairy cattle has long been treated as a background number in breeding records, a date on a calendar between conception and calving. A new genome-wide association study of German Holstein cows now shows that this number is anything but incidental. Researchers at the University of Hohenheim, together with colleagues at IT Solutions for Animal Production (vit) in Verden and the University of Göttingen, scanned the genomes of 34,497 German Holstein cows and identified specific chromosomal regions linked to how long a cow carries her calf. Their findings, published in Archives Animal Breeding, connect gestation length to some of the most economically and ethically charged traits in dairy farming: stillbirth and calving ease.
The study, led by Maximilian Zölch, is notable for its technical design. Earlier genome-wide association studies of gestation length in Canada, France, and Ireland either used window-based approaches, relied on lower-density 50K SNP chips, or included only sires, which made it impossible to separate the calf’s own genetic influence on gestation length from the mother’s. By contrast, the German team worked with imputed high-density genotypes: the animals were originally genotyped on 50K or low-density chips, and their data were imputed up to high density, yielding 479,759 SNPs per animal after quality filtering for minor allele frequency and deviations from Hardy-Weinberg equilibrium. Positions were mapped against the ARS-UCD1.2 reference genome assembly.
Crucially, the researchers analyzed gestation length as four distinct traits. The direct calf effect on gestation length was assessed separately for first and second parity, labeled GL_d1 and GL_d2, capturing the calf’s own genes that influence how long it is carried. The maternal effect, reflecting the cow’s genetic contribution to her own gestation length, was likewise split into first and second parity as GL_m1 and GL_m2. These traits were derived from deregressed breeding values estimated by vit using a single-step SNP BLUP model, with variance components drawn from Australian estimates by Haile-Mariam and Pryce, since German-specific estimates were not available. The number of observations ranged from 6,141 for the second-parity direct effect to 31,007 for the first-parity maternal effect.
The statistical machinery behind the association test was a mixed linear model implemented in GCTA. Each SNP was tested as a fixed effect while a random polygenic effect, built from a genetic relationship matrix constructed from the high-density genotypes, absorbed the background relatedness among the cows. This mixed-model approach is considered conservative because it models the candidate SNP both as a fixed and a random effect, but the team checked for distortion using the genomic inflation factor, which came out close to one for all four traits. That result indicates little to no false-positive inflation, giving the researchers confidence that their significant hits reflect genuine associations rather than artifacts of population structure. Two significance thresholds were applied: a Bonferroni-corrected genome-wide threshold and a more permissive suggestive threshold.
The results were striking in their asymmetry. For the direct calf effect in first parity, significant variants appeared on bovine chromosomes 14, 18, and 19, with 15 additional significant SNPs on chromosome 18 and 29 more on chromosome 19 beyond those found for second parity. For the direct effect in second parity, significant variants clustered on chromosomes 18 and 19, with a single variant on chromosome 18 and 41 on chromosome 19. The maternal effect told a much quieter story: only one variant, on chromosome 19, reached genome-wide significance, and that only for second parity. No significant associations emerged at all for the first-parity maternal effect.
This asymmetry makes biological sense when set against heritability estimates. The direct effect of the calf on gestation length is moderately heritable, estimated at 0.28 in heifers and 0.36 in cows, whereas the maternal heritability is a mere 0.04. In other words, the calf’s genome exerts far more control over the length of its own gestation than the mother’s does. The stronger signal for first-parity direct effects, despite the somewhat higher heritability reported for second parity, is most plausibly explained by sample size: the number of deregressed proofs available for second parity was more than 50 percent smaller than for first parity. The authors note that a larger second-parity dataset could reveal whether a genuine genomic difference exists between the two parities.
To interpret their hits, the researchers searched the Cattle QTL Database for known quantitative trait loci within 25 kilobases of each significant SNP, a window chosen because previous work on German Holsteins showed peak linkage disequilibrium at exactly this distance. The annotation exercise proved rich. Around the significant variants, the database listed QTLs for a suite of reproduction traits, with calving ease, stillbirth, and age at first calving the most frequently represented. Five previously reported QTLs for gestation length itself also sat adjacent to variants associated with the first-parity direct effect. Significantly more QTLs clustered around the first-parity hits than around the second-parity hits, mirroring the greater number of significant SNPs for that trait.
One association stands out for its mechanistic intrigue. A significant SNP for the first-parity direct effect, rs136803819 on chromosome 18, lies just 888 base pairs from a marker originally reported by Cole and colleagues in 2009, and both fall within the LOC618463 gene, which codes for sialic acid-binding Ig-like lectin 5, or SIGLEC5. Cole’s team proposed that elevated SIGLEC5 activity reduces the pool of free leptin, the hormone intimately involved in energy balance and pregnancy timing, and that this reduction contributes to longer gestation. The hypothesis fits neatly with later findings: a Polish study reported effects of a SIGLEC5 polymorphism on calving interval in Holstein-Friesian cattle, and human data have linked serum leptin levels during mid-to-late pregnancy to the initiation and progression of labor. A single gene thus threads together cattle breeding records and human obstetrics.
The broader pattern emerging from the study is that genomic regions governing gestation length overlap heavily with loci influencing other calving traits. Chromosome 18 harbors QTLs for calving ability, pregnancy rate, calf size, and birth index, while chromosome 19 carries QTLs for age at first calving. Stillbirth QTLs on both chromosomes sit close to the significant SNPs identified here. This overlap corroborates the team’s own earlier work showing significant global and local genetic correlations between gestation length and stillbirth, including locally segmented correlations pointing in opposite directions, a nuance that could be exploited in breeding programs aiming to reduce stillbirth without inadvertently lengthening gestations into the danger zone for dystocia.
The practical implications could be significant. Gestation length currently plays no role in routine breeding value estimation for German Holsteins, but the authors argue it should be incorporated into genomic selection, with emphasis on the direct calf effect, where the genetic potential for change is greatest. Gestation length occupies an intermediate optimum: too short a gestation compromises calf viability, while too long a gestation inflates birth weight and the risk of difficult calvings, and by the measure of perinatal mortality the shortened end of the distribution is the more dangerous one. The identified genomic regions could also serve as filters in subsequent gene expression analyses, narrowing candidate lists before expensive functional work begins. For a trait that sits at the intersection of fertility, animal welfare, and farm economics, the German Holstein genome has just yielded a considerably more detailed map.
Subject of Research: Genetic variants associated with gestation length in German Holstein dairy cattle
Article Title: Genome-wide association study of gestation length in German Holsteins
Article References: Zölch, M., Haas, V. P., Schmidtmann, C., Alkhoder, H., Tetens, J., & Bennewitz, J. (2026). Genome-wide association study of gestation length in German Holsteins. Archives Animal Breeding, 69(4), 563-568. https://doi.org/10.5194/aab-69-563-2026
Image Credits: AI Generated
Keywords: gestation length, German Holstein, genome-wide association study, GWAS, dairy cattle, stillbirth, calving ease, SIGLEC5, quantitative trait loci, genomic selection, cattle genetics, fertility traits
