AGD and AFC GO terms
Using early-in-life markers of reproductive characteristics may enhance the speed and success of genetic improvement in fertility. We investigated two phenotypes that can be measured early-in-life and are moderately heritable, to determine their association with traditional measures of reproductive success and genetic variation under a seasonal calving, pasture-based system. Cows were bred to be divergent in the New Zealand Fertility Breeding Value which estimates genetic merit for fertility. Cows consisted of two groups with an average positive (+ 5%) or a negative (- 5%) genetic merit for fertility traits and were expected to have large diversity in reproductive outcomes. Calves were genotyped at 41 d ± 3.1d of age (n = 538) and antral follicle counts (AFC) were measured when they were post-pubertal heifers before their first breeding (416 ± 15 d old; 92 d range; n = 520). The anogenital distance (AGD) was measured in 478 primiparous cows of this same population 50 - 60 d after the breeding start date when they were 881 ± 25 d old (145 d range). The AGD was shorter in animals with a positive genetic merit for fertility traits (based on parent averages). An indicator of herd reproductive success in a seasonal calving system (re-calving by 6 wk in lactation 2) was chosen for logistic regression with cross-validation, and if significant, a cut-off was calculated that categorized animals into groups. Both linear and quadratic regression was undertaken and the model with the greatest sensitivity for detection of non-pregnant cows used. The AGD linear model was significant with a sensitivity of 64% and a specificity of 48%. This model resulted in a cut-off of 102 mm which we used to classify cows as short (≤ 102 mm) or long (>102 mm) AGD animals. Primiparous cows with a short AGD were more likely to conceive within the first 3 and 6 wk of mating, and become pregnant as a primiparous cow, than those with a long AGD. The time from calving to conception was 20 d earlier in short AGD compared with long AGD cows. None of the models tested for AFC were significant, therefore cows were categorized into 3 groups based on previous work in seasonal systems. However, associations between fertility phenotypes and AFC group were limited. Genomic regions of interest for AGD and AFC did not overlap, indicating phenotypes were genetically independent. Overall, AGD appears as a promising early marker of fertility in seasonal grazing systems. Supplementary Table 1. Top 5 Gene Ontology (GO) terms associated with antral follicle count (AFC) or anogenital distance (AGD).
Steps to reproduce
Pathway analysis used MAGMA (version 1.07; de Leeuw et al., 2015) and followed 3 steps. Firstly, where possible, SNPs were annotated to the nearest gene using a boundary extended by 7 kb outside the transcription start and end sites of the gene, using bovine gene location (UMD3.1) from ensemble-biomaRt (www.ensembl.org/biomart). This annotation pipeline assigned 65,448 SNPs and resulted in a total of 22,429 genes being represented by one or more SNP. Secondly, a gene-based analysis step using P-values from the GWAS GEMMA output and thirdly, a gene-set analysis to determine associated pathways with AFC and AGD using the C2: curated gene sets or the C5: Gene Ontology (GO) gene sets available from MSigDB Collections (https://www.gsea-msigdb.org/gsea/msigdb/collections.jsp). All analyses in MAGMA are structured as a linear regression model on gene-level data. Z= β_0+ β_1 "G" _1+e Where Z was the phenotype vector, Gene-sets G1 were binary indicator variables, coded with “1” for genes in the gene-set, and with “0” otherwise; e was the residual vector. The intercept β0 represents the mean, and β1 the association specific to the gene-set 1.