Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2026 Mar 3;57(2):e70071. doi: 10.1002/age.70071

Integrative analysis of quantitative trait loci of alternative polyadenylation and GWAS highlights key regulators of milk production and immune traits in Holstein

Xinyue Tao 1, Yongjie Tang 1, Xueqin Liu 1, Miaozhuo Ni 1, Wanyi Lai 1, Siqian Chen 1, Ying Yu 1,✉
PMCID: PMC12955697  PMID: 41773938

Abstract

Understanding the genetic regulation of complex traits like mastitis in Holstein dairy cows remains a challenge, despite significant advancements in breeding techniques. This study integrates 3′‐untranslated region alternative polyadenylation (APA) quantitative trait loci (3′aQTL) mapping with genome‐wide association studies (GWASs) to identify key genetic variants and regulatory genes influencing traits during early lactation. Heritability analysis showed that, on average, cis‐acting SNPs explained 22% of percentage of distal polyA site usage index (PDUI) variance, supporting a genetic basis for APA regulation. Building on this, we identified 1143 significant cis‐3′aQTL associated with PDUI variation across 215 genes (3′aGenes), with significant enrichment in immune system and disease‐related pathways. Comparisons with expression quantitative trait loci (eQTL) revealed limited overlap, suggesting distinct regulatory mechanisms. SNP rs109421300 was found to co‐localize with GWAS signals for milk protein, mastitis and ketosis traits. Further analysis linked rs109421300 to the regulation of GRINA via APA, with predicted transcription factor interactions involving NRF1 and RUNX. These findings enhance our understanding of the genetic basis of economically important traits in dairy cows and provide a foundation for precision breeding strategies.

Keywords: alternative polyadenylation, complex traits, early lactation, Holstein, quantitative trait loci

INTRODUCTION

The lactation period is a critically important stage in the life cycle of dairy cows, profoundly affecting the nutrition and immune protection of their offspring as well as the distribution of their own energy needs (Bruckmaier & Gross, 2017). However, owing to the high energy demands of milk production, dairy cows frequently experience an energy deficit in the early lactation and peak milk production stages, impacting their health and reproductive traits and highlighting the need for research into the complex genetic mechanisms underlying the lactation process (Allen, 1996; Strucken et al., 2015). Despite significant advances in understanding the physiological aspects of the lactation cycle (Capuco & Akers, 2009), the genetic foundation and regulatory networks remain a challenge, particularly regarding the dynamic changes in genetic control of milk production traits during specific stages of lactation. The heritability of milk production traits typically ranges from moderate to high (e.g. 0.1–0.5 for milk yield, 0.1–0.3 for protein yield and 0.05–0.2 for somatic cell score), underscoring the substantial genetic contribution to these economically important traits (Liu, Xu, et al., 2022; Xue et al., 2022). Research in the early lactation period is crucial not only for identifying key genetic factors that affect dairy cow productivity but also for offering new strategies for disease prevention and management.

In recent years, research on genome‐wide association studies (GWASs) and the application of multiomics methods have provided a new perspective for revealing the complex genetic mechanisms behind different traits in cows. However, as most identified molecular markers are located in non‐coding regions, their regulatory characteristics have increasingly become a focus of attention (Xiang et al., 2018). Alternative polyadenylation (APA) is a key post‐transcriptional regulatory mechanism that generates mRNA isoforms with different 3′untranslated region (3′UTR) lengths by utilizing alternative polyadenylation sites, thereby influencing mRNA stability, localization and translation efficiency (Tian & Manley, 2017). Notably, studies of APA and APA‐associated quantitative trait loci have made important advances in humans (Li et al., 2021, 2022) but have yet to be reported in dairy cows, pointing to a new direction for research into complex traits.

To investigate genetic regulatory mechanisms in early lactation dairy cows, we extracted total RNA from peripheral blood leukocytes. Peripheral blood leukocytes serve as an accessible proxy for immune activity, which is particularly relevant in early lactation when cows are highly susceptible to infectious diseases such as mastitis and metabolic stress (Buggiotti et al., 2021; Wathes et al., 2021). Therefore, in this study we aimed to identify cis‐3′‐untranslated region APA quantitative trait loci (3′aQTL) and investigate their functional characteristics. By integrating GWAS data with cis‐3′aQTL information for co‐localization, we aimed to identify potential causal variants influencing complex traits during early lactation. This study provides new data for improving the accuracy of genomic selection for related traits in dairy cows.

MATERIALS AND METHODS

Collection of blood samples and milk samples

Our research team collected tail vein blood samples from 101 Chinese Holstein cows, with the sampling criteria being lactation days not exceeding 150 days and cows having no more than three parities. Among these cows, 56 were in their first parity, 17 in the second and 28 in the third. The days in milk of the sampled cows averaged 87 ± 30 days. The collected blood samples were stored in anticoagulant tubes for subsequent DNA and RNA extraction. All procedures involving experimental animals were approved by the Animal Welfare Committee of the China Agricultural University, Beijing, China. All efforts were made to minimize suffering and discomfort of the experimental animals.

RNA extraction, sequencing and quality control

For transcriptome analysis, we isolated the buffy coat from fresh anticoagulated blood samples by centrifugation at 3500 rpm for 15 min at room temperature. Total RNA was extracted from buffy coat using TRIzol reagent (Thermo Fisher Scientific, Waltham, MA, USA), following standard protocols. After centrifugation, the buffy coat was mixed with TRIzol, vortexed and homogenized. RNA was then purified through chloroform extraction, isopropanol precipitation and ethanol washing (Chomczynski & Sacchi, 2006). The extracted RNA was quantified using a Nano Drop spectrophotometer and assessed for quality using a 1% agarose gel. The integrity of RNA integrity number (RIN) was evaluated using an Agilent 2100 Bioanalyzer, and quantification was performed with a Qubit 2.0 Fluorometer. The RNA was then subjected to sequencing on an Illumina NovaSeq‐6000 platform, producing 150 bp paired‐end reads. Quality control of raw reads was performed using Trimmomatic (v0.39) with the following parameters: adapters – TruSeq3‐PE.fa:2:30:10, LEADING:3, TRAILING:3, SLIDINGWINDOW:4:15 and MINLEN:36 (Tang et al., 2024).

DNA genotyping and quality control

DNA was extracted from whole blood samples using the TIANamp Blood DNA Kit (Tiangen Biotech, Beijing, China) according to the manufacturer's protocol. DNA concentration and purity were assessed using a NanoDrop spectrophotometer, and integrity was confirmed by 1% agarose gel electrophoresis. Qualified DNA samples were then genotyped using the GGP Bovine 150 K BeadChip (Neogen, Lansing, MI, USA). Quality control measures included removing SNPs with a minor allele frequency < 0.05, SNPs significantly deviating from Hardy–Weinberg equilibrium (p < 1 × 10−5), variants with missing genotype rates >5% and samples with missing genotype rates >10%. SNP coordinates from the 150 K BeadChip were mapped from UMD3.1 to ARS‐UCD1.2 based on rsid obtained from dbSNP (Sherry et al., 2001). After quality control, 101 individuals and 95 799 SNPs remained for further analysis.

Quantification of gene expression and APA usage

Gene expression quantification was performed following previously published pipelines in cattle (Liu, Gao, et al., 2022; Tang et al., 2024). High‐quality RNA‐seq reads were aligned to the ARS.UCD.1.2 reference genome (Rosen et al., 2020) using star (v2.7.9a) (Dobin et al., 2013). Gene expression levels were estimated using stringtie (v2.1.5) (Pertea et al., 2016) to obtain transcripts per million (TPM) values, and featurecounts (v2.0.2, subread package) (Liao et al., 2014) was employed to quantify transcript counts. To ensure data robustness, only genes with TPM values greater than 0.1 in at least 20% of samples and those with read counts higher than 6 in 20% or more of the samples were included for further analysis. Normalization of gene expression data was performed using the TMM (trimmed mean of M‐values) method in edger (v3.34.1) (Robinson et al., 2010).

The number of APA events was assessed using DaPars2 (Feng et al., 2018), which estimates the percentage of distal polyA site usage index (PDUI) from RNA‐seq data. Using the ensembl gene annotation for the bovine ARS‐UCD1.2 reference genome, we first generated transcript‐level 3′UTR annotations with the script DaPars_Extract_Anno.py from the DaPars2 tool package. This script processes a gene annotation file to systematically identify 3′UTRs. It locates the stop codon for each transcript isoform and defines the downstream region as the transcript‐specific 3′UTR. To generate a canonical set for analysis, it merges the 3′UTRs of all isoforms for each gene, outputting a BED file with the consolidated genomic coordinates. RNA‐seq data from multiple samples were processed jointly, and sequencing depth was normalized using samtools (v1.9) (Li & Durbin, 2009). dapars2 was then applied to detect de novo APA sites and calculate PDUI values for each transcript. The PDUI ranges from 0 to 1, with higher values indicating greater usage of distal polyA sites and therefore longer 3′UTRs, while lower values reflect preferential usage of proximal sites and shorter 3′UTRs (Feng et al., 2018).

Analysis of the covariates in the QTL identification model

Confounding factors and population genetic structure can affect the identification of cis‐3′aQTL; therefore, covariates are introduced to correct for these effects. Initially, snprelate (v1.26.0) (Zheng et al., 2012) was utilized to calculate the genotype principal components (PCs) from genotype data, with the first five genotype PCs being included as covariates. To identify biological and technical factors influencing APA events, we performed principal component analysis on the normalized PDUI matrix. The top 10 PCs of PDUI variation were then correlated with sample metadata (such as herd, RIN, parity) using Pearson's correlation. Afterwards, lactation weeks, parity and RIN were included as the main factors affecting PDUI variation in the covariates.

Additionally, the peer v1.3 (R package) software was used to identify unknown confounding factors caused by unmeasured reasons through the probabilistic estimation of expression residuals (peer) method (Stegle et al., 2010). Normalized PDUI values were input to identify peer factors, and based on the impact of different numbers of peer factors on the quantity of identified 3′aQTL‐3′aGene pairs, an appropriate number of peer factors were selected to be included in the model as covariates.

Estimation of heritability

The cis‐SNPs used to estimate the heritability of PDUI were defined as variants located within ±1 Mb of the 3′UTR of each target gene. GCTA (v1.93.3 beta2) (Yang et al., 2011) was used to construct a genetic relationship matrix based on the cis‐SNPs for each gene. Subsequently, heritability of PDUI was estimated using a linear mixed model under the restricted maximum likelihood framework, while correcting for the aforementioned covariates. The significance of the genetic component was evaluated using a likelihood ratio test, and only genes with significant heritability estimates (p < 0.05) were retained.

Identification of the cis‐3′aQTL

The PDUI computed by the dapars2.0 algorithm serves as the phenotype for identifying cis‐3′aQTL, with SNP genotyping data as the independent variable. A 1 Mb window upstream and downstream of the 3′UTR region was selected for identification through linear regression analysis conducted by Matrix eQTL (Shabalin, 2012), incorporating five genotype PCs, parity, weeks of lactation, RIN and the first two peer factors as covariates. Multiple testing correction was performed using the Benjamini–Hochberg procedure. Only SNP–gene pairs with a false discovery rate (FDR) < 0.05 were defined as significant cis‐3’aQTLs. Subsequently, the corresponding host genes harboring at least one significant cis‐3’aQTL were defined as 3’aGenes.

Prediction of cis‐3′aQTL effect

To explore the genomic distribution and regulatory features of cis‐3′aQTL, we first plotted their positions across the genome, with distances to transcription termination sites on the x‐axis. We then used the Ensembl variant effect predictor (VEP) to annotate these cis‐3′aQTL, providing insights into their impacts on transcripts, proteins and regulatory regions.

To assess significant enrichment of cis‐3′aQTL in specific genomic regions, we conducted Fisher's exact test. This test compared the distribution of cis‐3′aQTL with all SNP loci on the genotyping array, identifying regions where cis‐3′aQTL are significantly enriched.

Overlap and comparative analysis of cis‐3′aQTL and previously identified cis‐eQTL

To elucidate the regulatory characteristics of cis‐3′aQTL independent of traditional cis‐eQTL, we compared the cis‐3′aQTL identified in this study with cis‐ eQTL previously identified using the same dataset by our team (Tang et al., 2024). Briefly, this dataset consisted of RNA‐seq data from peripheral blood leukocytes and DNA genotypes from the GGP Bovine 150 K BeadChip, collected from 101 Chinese Holstein cows in early lactation. The cis‐eQTL were mapped using Matrix eQTL, adjusting for genetic PCs and relevant covariates. This comparison aimed to isolate both common and unique SNPs and genes, to further investigate their functional properties. For visual representation of the overlap and distinctions between these genetic variants, we utilized the VennDiagram package in R to generate Venn diagrams. We also conducted a comparative analysis of effect sizes between the 3′aQTL and eQTL, generating scatter plots to visualize the correlation. Subsequently, the identified shared and unique genes were subjected to KOBAS (http://bioinfo.org/kobas) for GO (Gene Ontology) and KEGG (Kyoto Encyclopedia of Genes and Genomes) enrichment analyses, with the species selected as vertebrate/Bos taurus (cow). Significant pathways (p < 0.05) were identified, and the top 20 biological functions and pathways were visualized in a bubble chart to highlight the key functional categories and pathways.

We selected an immune pathway related to immune regulation in mastitis, which was enriched among the specific 3′aGenes, and conducted interaction network analysis on the enriched genes in the pathway using STRING (Szklarczyk et al., 2023). In addition, we visualized 3′UTR usage differences for representative genes with the most significant cis‐3′aQTL by plotting PDUI values and inspecting read coverage profiles in the integrative genomics viewer.

Colocalization of cis‐3′aQTL loci with GWAS loci for complex traits

To examine whether cis‐3′aQTL overlap with significant loci identified in GWASs for complex traits in dairy cattle, we utilized GWAS summary statistics for milk protein, mastitis and ketosis from 27 214 bulls for colocalization analysis (Jiang et al., 2019). Colocalization analysis was performed using the coloc (v5.1.0) software (Giambartolomei et al., 2014), which applies a Bayesian framework to determine if SNPs associated with these traits share a causal relationship with cis‐3′aQTL. Genes were identified based on the posterior probability of colocalization (PP.H4). Specifically, PP.H4 > 0.70 in both cis‐3′aQTL and GWAS loci was used to determine colocalization. Prior to this analysis, cis‐3′aQTL were converted to the UMD3.1 genome build to ensure consistency with the GWAS reference genome version. To further ensure the robustness of the analysis, the cis‐3′aQTL data were first split by single gene into individual files. These split files were then merged with the corresponding GWAS data files. This approach allowed for detailed colocalization analysis for each gene separately. Manhattan plots illustrating the results were generated using the qqman (v0.1.8) package (Turner, 2018).

Next, we used homer (Heinz et al., 2010) to predict transcription factor binding motifs in the ±50 bp of the colocalization SNP. We then analyzed the expression levels of these predicted transcription factors in relation to alternative polyadenylation site usage, aiming to elucidate the regulatory mechanism of the cis‐3′aQTL. This approach helps to clarify how transcription factor binding may influence APA events associated with the cis‐3′aQTL.

RESULTS

Covariate selection and heritability estimation in the cis‐3′aQTL identification

To identify cis‐3′aQTL, transcriptome sequencing data was used to assess gene expression and 3′‐end alternative polyadenylation. Since transcriptomic data are sensitive to confounding factors such as batch effects, biological variations and technical issues (Walker et al., 2020), we conducted correlation analyses to address these influences. Specifically, we examined the relationships between the first 10 PCs of PDUI and known factors: lactation week, parity and RIN. Our results revealed significant correlations between these factors and PCs of PDUI (Figure 1a). For the significance testing of Pearson correlations, we found that each factor showed significant correlations (p < 0.05) with at least two of the first 10 PCs (Table S1). As a result, these three factors were included as covariates in our cis‐3′aQTL identification model.

FIGURE 1.

FIGURE 1

Open in a new tab

Factors influencing cis‐3′‐untranslated region alternative polyadenylation (APA) quantitative trait loci (3′aQTL) identification and heritability distribution. (a) Correlation between the first 10 principal components (PCs) of polyA site usage index (PDUI) and weeks of lactation, parity and RNA integrity number (RIN) values; size of the point corresponds to the absolute value of the correlation coefficient. (b) The number of peer factors used and their corresponding number of identified 3′aQTL‐3′a gene pairs. (c) PDUI as the heritability estimate of the phenotype.

Additionally, to account for unknown confounding factors, we used the peer software, which is based on a Bayesian framework. We tested models with varying numbers of peer factors (0–15) and found that using two peer factors resulted in the identification of the most cis‐3′aQTL (Figure 1b). Consequently, two peer factors were selected for inclusion in the final model used for cis‐3′aQTL identification.

Specifically, we estimated the heritability of PDUI variation, considering all cis‐SNPs within ±1 Mb of the 3′UTR of each gene. On average, cis‐SNPs explained 22.0% of the genetic variation in PDUI, considering only genes with significant heritability estimates (p < 0.05). This indicates that genetic variation in cis regions contributes measurably to 3′UTR differences, and the estimated heritability is comparable with that reported in human tissues (Li et al., 2021). The top 15 genes with the highest heritability of PDUI included TXNIP, ERCC2, UBC, TBC1D8 and POLR1H (Table S2). Their high heritability indicates strong genetic contributions to 3′UTR usage variation and underscores their potential roles in APA regulation.

Identified cis‐3′aQTL in early lactation

In this study, the Matrix eQTL software was utilized to identify cis‐3′aQTL within a window region of ±1 Mb around the 3′UTR, taking into account both known and inferred unknown covariates (five genotype PCs, lactation weeks, parity, RIN and two peer factors) for adjustment, aiming to pinpoint genetic variants associated with differential 3′UTR expression. We identified 1143 cis‐3′aQTL‐3′UTR associations (FDR < 0.05), along with 215 genes associated with 3′aQTL (3′aGene). Table 1 lists the top 10 most significant identified cis‐3′aQTL and their corresponding 3′aGenes. It can also be seen that a gene may be associated with more than one cis‐3′aQTL, e.g. the POLR1H, ENSBTAG00000019018 and ENSBTAG00000050510 genes in the table are all associated with more than one cis‐3′aQTL.

TABLE 1.

Top 10 most significant cis‐3′aQTL and their corresponding 3′aGenes.

cis‐3′aQTL 3′aGene p‐value Beta
rs109675766 ENSBTAG00000019018 3.06 × 10−24 1.16
rs134596536 ENSBTAG00000055240 9.33 × 10−24 −1.07
rs110027095 UBC 1.92 × 10−21 1.31
rs110730562 ENSBTAG00000019018 8.47 × 10−21 1.42
rs41892940 ENSBTAG00000050510 9.22 × 10−21 1.39
rs109012807 ENSBTAG00000050510 9.22 × 10−21 1.39
rs110763614 ENSBTAG00000019018 1.22 × 10−18 1.33
rs109318615 ENSBTAG00000007075 1.47 × 10−18 −1.07
rs134294517 POLR1H 1.82 × 10−18 1.09
rs137214390 POLR1H 3.02 × 10−17 1.05
Open in a new tab

Regulatory characteristics of the cis‐3′aQTL in early lactation

To explore whether the early lactation cis‐3′aQTL in dairy cows are associated with regulatory regions of the genome, the spatial location of cis‐3′aQTL was explored. The results indicate that the distribution of identified cis‐3′aQTL across the genome resembles a normal distribution, predominantly located near transcription termination sites. As the physical distance from the transcription termination site increases, the number of cis‐3′aQTL gradually decreases (Figure 2a).

FIGURE 2.

FIGURE 2

Open in a new tab

Genomic features of cis‐3′‐untranslated region alternative polyadenylation (APA) quantitative trait loci (3′aQTL). (a) Distribution of cis‐3′aQTL distance to transcription termination sites. (b) Enrichment (Fisher's exact test) of cis‐3′aQTL with Ensembl variant effect predictor (vep)‐predicted SNP effects, respectively. The point and error bars indicate the odds ratio and 95% CI; the asterisk indicates significance (p < 0.05).

Enrichment analysis (Fisher's exact test) was conducted to compare the distribution of cis‐3′aQTL across gene functional regions as predicted by variant effect predictor (vep) relative to all SNPs on the SNP chip. Notably, variants located in the 3′UTR (OR = 3.49, 95% CI: 2.16–5.36, p = 1.27 × 10−6) and 5′UTR (OR = 3.24, 95% CI: 1.03–7.78, p = 0.022) regions displayed increased odds of enrichment. Downstream (OR = 2.92, 95% CI: 2.42–3.51, p < 2.2 × 10−16) and upstream (OR = 2.47, 95% CI: 2.03–2.98, p < 2.2 × 10−16) variants were also significantly enriched. Conversely, intergenic regions demonstrated reduced odds (OR = 0.33, 95% CI: 0.29–0.38, p < 2.2 × 10−16) (Figure 2b). These results indicate that cis‐3′aQTL are more likely than genotyped SNPs to occur in gene‐proximal regions.

Similarities and differences between cis‐3′aQTL and cis‐eQTL

To examine the shared and unique genetic variations between cis‐3′aQTL and eQTL, we conducted a Venn analysis. Applying a filtering criterion of p < 0.01, we identified a total of 1718 shared SNPs, alongside 7256 eQTL‐specific SNPs and 6881 cis‐3′aQTL‐specific SNPs (Figure 3a). Furthermore, the analysis revealed 1026 shared genes, 4929 eGene‐specific genes and 1340 3′aGene‐specific genes (Figure 3b), indicating that the majority of cis‐3′aQTL are independent from eQTL and affect phenotypic traits in different ways. Next, we explored the relationship between eQTL and cis‐3′aQTL by analyzing the effect sizes of paired gene loci (eQTL_eGene and 3′aQTL_3aGene, p < 0.01). The linear regression analysis showed a positive slope, suggesting that in some cases, eQTL and cis‐3′aQTL effect sizes may align in the same direction. However, the correlation analysis yielded an R 2 value of 0.0342 (p = 0.0745, Figure 3c), indicating a relatively weak correlation between the effect sizes of the two types of QTL and not significant. This modest correlation suggests that, while there may be instances of shared influence, a significant portion of the effects of eQTL and cis‐3′aQTL operate independently. The majority of their effects on gene regulation appear to be distinct, emphasizing the complexity of their regulatory mechanisms.

FIGURE 3.

FIGURE 3

Open in a new tab

Comparison of expression quantitative trait loci (eQTL) and cis‐3′‐untranslated region alternative polyadenylation (APA) quantitative trait loci (3′aQTL) effects, enrichment analysis and functional insights into Th17 pathway genes. (a) 3′aQTL and eQTL sharing and specific SNPs. (b) 3′aGene and eGene sharing and specific Genes (c). Scatterplot of effect size of 3′aQTL with eQTL. (d) Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis of specific 3′aGenes. (e) KEGG enrichment analysis of shared 3′aGenes and eGenes. (f) Interaction network of genes in the Th17 cell differentiation pathway enriched for specific 3′aGenes. (g) Genotype‐dependent variation in percentage of polyA site usage index (PDUI) for BoLA‐DMA. (h) Genotype‐specific 3′UTR structure visualization for BoLA‐DMA; the box represents the 3′UTR of the transcript.

Functional enrichment analysis of 3′aGene‐specific genes and shared genes

We performed functional enrichment analysis for both 3′aGene‐specific and shared genes. Overall, both 3′aGene‐specific and shared genes were enriched in immune and infection‐related pathways. However, the 1340 3′aGene‐specific genes showed relatively focused enrichment in immune response and signaling pathways (e.g. ‘Th17 cell differentiation’, ‘Epstein–Barr virus infection’ and ‘Colorectal cancer’), indicating their potential involvement in specialized immune regulatory processes (Figure 3d). In contrast, the 1026 shared genes exhibited broader enrichment across fundamental cellular and metabolic processes, including ‘Bacterial invasion’, ‘Endocytosis’ and ‘Protein processing in the endoplasmic reticulum’ (Figure 3e). These results indicate that while 3′aGene‐specific loci may fine‐tune immune‐related gene regulation through post‐transcriptional mechanisms, shared loci probably contribute to more general biological processes that support core cellular functions.

3′aGene‐specific genes regulation in the Th17 cell differentiation pathway

The Th17 cell differentiation pathway, which plays a critical role in immune regulation in mastitis (Cebron et al., 2020), was found to be significantly enriched among specific 3′aGenes. To investigate potential regulatory mechanisms, we constructed a gene interaction network using the genes involved in this pathway (Figure 3f). The network shows several key transcription factors and immune‐related genes, including BoLA family genes, TYK2 and STAT5b; these BoLA genes, along with other immune regulators like TYK2 and STAT5b, form a distinct subnetwork, highlighting their potential coordinated roles in immune modulation.

To further understand the functional impact of these specific 3′aGenes, we examined the PDUI variation across different genotypes (AA, AG and GG) for the BoLA‐DMA gene, a key gene in the Th17 cell differentiation pathway associated with the most significant cis‐3′aQTL. The results reveal a clear genotype‐dependent pattern (Figure 3g), suggesting that this cis‐3′aQTL may influence the polyadenylation pattern of BoLA‐DMA, potentially affecting its stability or translation efficiency. The 3′UTR structure of BoLA‐DMA varies across genotypes, with distinct 3′UTR lengths and coverage corresponding to each genotype (Figure 3h), highlighting genotype‐specific differences in APA usage. These findings support the idea that 3′aQTL may play a crucial role in modulating immune gene expression through polyadenylation mechanisms, suggesting a pathway‐specific fine‐tuning process relevant to immune‐related traits.

Co‐localization of cis‐3′aQTL and complex trait GWAS signals

Alternative polyadenylation is widely present in mammals and plays an important role in shaping transcript isoforms and gene expression regulation (Li et al., 2021). Therefore, in order to analyze whether the identified cis‐3′aQTL in this study contribute to explaining the genetic mechanism of significant GWAS loci for complex traits in cattle, this study conducted co‐localization analysis between the identified 1143 cis‐3′aQTL‐3′aGene pairs and milk protein as well as two health traits of mastitis and ketosis. The results identified 97 SNPs as shared colocalization signals across three traits with a posterior probability (PP.H4) greater than 0.7 (Table S3). Among these, rs109421300 had a PP.H4 of 1 and was the most significant SNP associated with milk protein, ketosis and mastitis traits (Figure 4a–c). The SNP impacts the expression of ENSBTAG00000000312 by regulating the APA usage of the 3′UTR of the ENSBTAT00000000410 transcript. Moreover, rs109421300 is also the most significant cis‐3′aQTL within the region spanning 14 000 000–24 000 000 bp on chromosome 14 (Figure 4d). This dual role as both a trait‐associated SNP and a cis‐3′aQTL underscores its potential pivotal regulatory function in APA and its broader impact on gene expression.

FIGURE 4.

FIGURE 4

Open in a new tab

Colocalization of rs109421300 with genome‐wide association study (GWAS) traits. (a–c) GWAS Manhattan plots of milk protein, ketosis and mastitis, respectively. (d) Manhattan plot of cis‐3′‐untranslated region alternative polyadenylation (APA) quantitative trait loci (3′aQTL).

Among the 97 colocalized SNPs, rs109421300 was the most significant across milk protein, mastitis and ketosis traits. Although this SNP is located within the DGAT1 gene region, our cis‐3′aQTL analysis indicated that it regulates APA of the GRINA transcript (ENSBTAT00000000410). Therefore, we focused on GRINA to investigate the potential functional consequences of rs109421300 on 3′UTR usage and downstream gene regulation.

rs109421300 modulates GRINA APA patterns and transcription factor abundance

To further explore the regulatory mechanisms of rs109421300, we analyzed the PDUI for GRINA across different genotypes of rs109421300 (Figure 5a). The results demonstrated a clear genotype‐dependent pattern, with significant differences in PDUI between AA, AG and GG genotypes. This finding supports the hypothesis that rs109421300 affects GRINA's APA usage. Then we performed transcription factor binding motif predictions within a 50 bp region surrounding rs109421300. The analysis identified NRF1 and RUNX as potential transcription factors binding near the SNP site (Figure 5b). Subsequently, we analyzed the correlation between GRINA's PDUI and the expression levels (TPM) of NRF1 and RUNX to investigate whether APA events in GRINA are associated with changes in transcription factor abundance. The results show a significant negative correlation between GRINA's PDUI and NRF1's expression (Figure 5c, r = −0.4483, p = 2.59 × 10−6) and a similarly significant negative correlation between GRINA's PDUI and RUNX's expression (Figure 5d, r = −0.3496, p = 0.0003). These findings suggest that the 3′UTR length of GRINA, as influenced by rs109421300, may play a role in modulating the expression levels of these transcription factors.

FIGURE 5.

FIGURE 5

Open in a new tab

Impact of rs109421300 on GRINA 3′untranslated region (3′UTR) alternative polyadenylation (APA) and its connection to NRF1 and RUNX expression. (a) Genotype‐dependent PDUI variation in GRINA (b) Match motif of NRF1 and RUNX (c‐d) Correlation between GRINA polyA site usage index (PDUI) and expression of NRF1 and RUNX.

DISCUSSION

Our study provides a comprehensive analysis of cis‐3′aQTL in dairy cows during early lactation, identifying variants associated with APA and potentially contributing to post‐transcriptional regulation of gene expression. The identified cis‐3′aQTL predominantly localize near transcription termination sites, suggesting their key role in post‐transcriptional regulation. This distribution pattern is consistent with previous findings in human tissues, highlighting the importance of 3′UTR regions in controlling gene expression (Li et al., 2021). The observed enrichment in 3′UTRs and other regulatory regions, such as the 5′UTR and upstream/downstream regions, underscores the importance of these variants in RNA processing, including alternative polyadenylation. Such regulatory variants are known to influence transcript stability and translation efficiency, thereby modulating gene expression (Gruber & Zavolan, 2019; Tian & Manley, 2017). It should be noted that the sample size of this study limits the detection of weaker‐effect cis‐3′aQTL; the subsequent genomic and functional analyses were conducted to assess the biological relevance of the identified loci.

Our comparative analysis demonstrated both unique and shared genetic loci between cis‐3′aQTL and cis‐eQTL. Venn analysis showed limited overlap in SNPs and genes, indicating that most cis‐3′aQTL function independently from cis‐eQTL. The relatively weak correlation between effect sizes of cis‐3′aQTL and cis‐eQTL further supports the idea that these two regulatory mechanisms operate through distinct pathways. These observations highlight the complementary roles of cis‐3′aQTL in post‐transcriptional regulation, compared with the more transcription‐focused role of cis‐eQTL. Similar patterns have been observed in human studies, where APA‐associated QTL showed only partial overlap with eQTL, suggesting that APA contributes to gene regulation largely through post‐transcriptional mechanisms rather than transcriptional variation (Shulman & Elkon, 2020).

The functional enrichment analysis of specific 3′aGenes indicates that cis‐3′aQTL may have targeted regulatory roles in pathways essential for immune function and disease response. The enrichment of the Th17 cell differentiation pathway among these 3′aGenes suggests that APA regulation may play a specialized role in modulating immune dynamics, particularly in response to inflammatory signals (Wang et al., 2023). Notably, the clustering of BoLA family genes with TYK2 and STAT5b in the Th17 pathway points to a coordinated regulatory network within immune‐related processes. The BoLA genes, part of the bovine leukocyte antigen family, are central to antigen presentation and immune recognition, playing a critical role in initiating and modulating immune responses (Behl et al., 2012). TYK2, a key signaling molecule in the JAK–STAT pathway, is crucial for cytokine signaling and immune cell differentiation, particularly in the activation of T‐helper cell pathways like Th17. STAT5b, a transcription factor activated by TYK2, further influences the development and function of immune cells by regulating genes associated with immune tolerance, inflammation and cellular proliferation (Seif et al., 2017). This clustering pattern implies that cis‐3′aQTL might not only influence gene expression at the individual gene level but could also drive broader network effects by collectively modulating the APA of genes involved in immune pathways. By impacting the APA of these interconnected genes, cis‐3′aQTL may fine‐tune the immune response, possibly enhancing immune tolerance and modulating inflammatory responses during early lactation.

By focusing on 3′UTR regulation and alternative polyadenylation, we identified novel mechanisms through which genetic variants influence milk composition and health traits. By integrating cis‐3′aQTL with GWAS signals, we uncovered 97 co‐localized SNPs with high posterior probabilities, suggesting shared genetic control over traits like milk protein and mastitis. The key SNP, rs109421300, plays a critical role in controlling the APA usage of the GRINA gene, which is known for its role in calcium regulation and metabolic processes (Jimenez‐Gonzalez et al., 2019). By regulating GRINA's APA patterns, rs109421300 may influence cellular calcium levels, apoptosis and metabolic pathways within mammary epithelial cells, impacting the synthesis and secretion of milk protein and fat. This highlights rs109421300 as a potential genetic marker for breeding programs aimed at enhancing early lactation performance and improving dairy cow resilience.

Further analysis of transcription factor binding motifs near rs109421300 identified NRF1 and RUNX as potential regulators. Correlation analyses showed significant negative associations between GRINA's PDUI and the expression levels of these transcription factors, suggesting that GRINA's 3′UTR could act as a regulatory hub. Longer 3′UTR variants may provide additional binding sites for repressive elements such as microRNAs or RNA‐binding proteins, leading to reduced mRNA stability or translation efficiency, whereas shorter variants may enhance transcript stability and transcription factor binding efficiency. Similar mechanisms have been observed in humans, where PAS‐disrupting SNPs induce 3′UTR lengthening accompanied by decreased gene expression and altered disease susceptibility (Shulman & Elkon, 2020). NRF1 plays a crucial role in mitochondrial biogenesis, maintaining mitochondrial DNA integrity and managing oxidative stress by regulating the expression of antioxidant proteins and genes involved in energy metabolism (Bugno et al., 2015; Park et al., 2016). This is particularly important in high‐energy‐demanding tissues like the mammary gland during lactation, where oxidative stress management is essential to maintain cellular health and function (Baumgard et al., 2017). RUNX, on the other hand, is a key transcription factor involved in immune cell differentiation, particularly in promoting the development of T‐cells, and it plays a significant role in controlling inflammatory responses. RUNX also participates in cytokine regulation and has been shown to mediate immune tolerance and inflammation control (Collins et al., 2009), both of which are critical for lactating dairy cows under metabolic stress. Thus, the influence of rs109421300 on GRINA's 3′UTR could modulate transcription factor activity, potentially affecting milk production and health traits by enhancing cellular resilience and optimizing immune function.

The SNP rs109421300, located within an intronic region of the DGAT1 gene, has been identified in our previous studies as an eQTL significantly associated with the differential expression of DGAT1 across its three genotypes (Tang et al., 2024). Additionally, rs109421300 shows strong co‐localization with GWAS signals linked to traits such as milk yield, fat content and mastitis, underscoring its importance in complex traits. Given that DGAT1 is a key gene in triglyceride synthesis (Grisart et al., 2004), the variation in DGAT1 expression driven by rs109421300 suggests that this SNP directly influences lipid metabolism and milk fat production through transcriptional regulation within DGAT1. The dual role of rs109421300 as an eQTL for DGAT1 and a 3′aQTL for GRINA highlights its pleiotropic regulatory potential, suggesting that it affects lactation traits through distinct yet interconnected mechanisms. These findings raise the possibility of a regulatory interaction between rs109421300's roles as a 3′aQTL and an eQTL. As a 3′aQTL, rs109421300 alters the APA usage of GRINA, potentially influencing its transcript stability or translation efficiency, which may have downstream effects on cellular calcium signaling and stress responses‐processes. Such changes in cellular metabolism could create feedback signals that indirectly influence DGAT1 transcription, given its central role in triglyceride synthesis. Conversely, shifts in DGAT1 expression and lipid biosynthesis, driven by rs109421300 as an eQTL, might also alter the cellular environment, potentially impacting GRINA's functional role in stress and calcium regulation. This interplay suggests a bidirectional regulatory network that warrants further investigation.

The dual regulatory role of rs109421300 demonstrates how a single SNP can integrate transcriptional and post‐transcriptional regulation to optimize metabolic and immune responses during lactation. This interconnected mechanism not only highlights the complexity of genetic regulation but also emphasizes the importance of combining eQTL and 3′aQTL analyses to capture the full spectrum of SNP effects on gene expression. To validate this hypothesis, functional experiments using CRISPR/Cas9 to manipulate rs109421300 in bovine mammary cell lines could assess its impact on DGAT1 expression, GRINA's APA patterns and related phenotypic traits under varying metabolic conditions. Additionally, transcriptomic and metabolomic profiling under stress conditions could provide further insights into how rs109421300 coordinates these pathways to optimize lactation performance.

CONCLUSIONS

The role of APA in transcript regulation is increasingly recognized as a crucial factor in gene expression. Our results demonstrate that cis‐3′aQTL play a significant role in APA usage, which in turn affects the stability and translation efficiency of key genes. This regulatory mechanism is particularly important in the context of early lactation, where the metabolic demands of milk production place unique pressures on gene regulation. Future studies exploring the functional effects of APA in dairy cows could provide further insights into how genetic variation impacts lactation performance.

AUTHOR CONTRIBUTIONS

Xinyue Tao: Conceptualization, formal analysis; investigation; methodology; validation; visualization; writing – original draft; writing – review and editing. Yongjie Tang: Conceptualization; Data Curation; software; formal analysis; investigation; methodology; validation; writing – review and editing. Xueqin Liu: Resources (sample collection and data generation). Miaozhuo Ni: Writing – review and editing. Wanyi Lai: Writing – review and editing. Siqian Chen: Writing – review and editing. Ying Yu: Conceptualization; project administration; supervision; writing – review and editing.

FUNDING INFORMATION

This study was financially supported by the National Key R&D Program of China (2021YFD1200903, 2023YFF1000902), the National Natural Science Foundation of China (32472864, 32302706), the National Natural Science Foundation of China ‐ Pakistan Science Foundation Joint Research Project (31961143009), the earmarked fund for CARS‐36, the Program for Changjiang Scholar and Innovation Research Team in University (IRT‐15R62) and the Seed Fund (CAU).

CONFLICT OF INTEREST STATEMENT

There is no conflict of interest with any financial organization regarding the materials discussed in the manuscript.

Supporting information

Appendix S1.

AGE-57-0-s001.docx (24.6KB, docx)

ACKNOWLEDGEMENTS

The authors thank the anonymous reviewers for constructive comments and suggestions on the manuscript, and all the members of the Animal Molecular and Quantitative Genetics Laboratory in China Agricultural University. This study was supported by High‐performance Computing Platform of China Agricultural University.

Tao, X. , Tang, Y. , Liu, X. , Ni, M. , Lai, W. , Chen, S. et al. (2026) Integrative analysis of quantitative trait loci of alternative polyadenylation and GWAS highlights key regulators of milk production and immune traits in Holstein. Animal Genetics, 57, e70071. Available from: 10.1002/age.70071

Xinyue Tao and Yongjie Tang are co‐first authors.

DATA AVAILABILITY STATEMENT

The raw RNA sequence data reported in this paper have been deposited in the Genome Sequence Archive in National Genomics Data Center, China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences (GSA: CRA012735). The public GWAS summary statistics from Figshare (https://figshare.com/s/ea726fa95a5bac158ac1). The SNP BeadChip data for the current study are available from the corresponding author on reasonable request.

REFERENCES

  1. Allen, M.S. (1996) Physical constraints on voluntary intake of forages by ruminants. Journal of Animal Science, 74, 3063–3075. [DOI] [PubMed] [Google Scholar]
  2. Baumgard, L.H. , Collier, R.J. & Bauman, D.E. (2017) A 100‐year review: regulation of nutrient partitioning to support lactation. Journal of Dairy Science, 100, 10353–10366. [DOI] [PubMed] [Google Scholar]
  3. Behl, J.D. , Verma, N.K. , Tyagi, N. , Mishra, P. , Behl, R. & Joshi, B.K. (2012) The major histocompatibility complex in bovines: a review. ISRN Veterinary Science, 2012, 872710. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bruckmaier, R.M. & Gross, J.J. (2017) Lactational challenges in transition dairy cows. Animal Production Science, 57, 1471–1481. [Google Scholar]
  5. Buggiotti, L. , Cheng, Z. , Salavati, M. , Wathes, C.D. & Genotype plus Environment C . (2021) Comparison of the transcriptome in circulating leukocytes in early lactation between primiparous and multiparous cows provides evidence for age‐related changes. BMC Genomics, 22, 693. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bugno, M. , Daniel, M. , Chepelev, N.L. & Willmore, W.G. (2015) Changing gears in Nrf1 research, from mechanisms of regulation to its role in disease and prevention. Biochimica et Biophysica Acta, 1849, 1260–1276. [DOI] [PubMed] [Google Scholar]
  7. Capuco, A.V. & Akers, R.M. (2009) The origin and evolution of lactation. Journal of Biology, 8, 37. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cebron, N. , Maman, S. , Walachowski, S. , Gausseres, B. , Cunha, P. , Rainard, P. et al. (2020) Th17‐related mammary immunity, but not a high systemic Th1 immune response is associated with protection against E. Coli mastitis. npj Vaccines, 5, 108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Chomczynski, P. & Sacchi, N. (2006) The single‐step method of RNA isolation by acid guanidinium thiocyanate‐phenol‐chloroform extraction: twenty‐something years on. Nature Protocols, 1, 581–585. [DOI] [PubMed] [Google Scholar]
  10. Collins, A. , Littman, D.R. & Taniuchi, I. (2009) RUNX proteins in transcription factor networks that regulate T‐cell lineage choice. Nature Reviews. Immunology, 9, 106–115. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Dobin, A. , Davis, C.A. , Schlesinger, F. , Drenkow, J. , Zaleski, C. , Jha, S. et al. (2013) STAR: ultrafast universal RNA‐seq aligner. Bioinformatics, 29, 15–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Feng, X. , Li, L. , Wagner, E.J. & Li, W. (2018) TC3A: the cancer 3′ UTR atlas. Nucleic Acids Research, 46, D1027–D1030. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Giambartolomei, C. , Vukcevic, D. , Schadt, E.E. , Franke, L. , Hingorani, A.D. , Wallace, C. et al. (2014) Bayesian test for colocalisation between pairs of genetic association studies using summary statistics. PLoS Genetics, 10(5), e1004383. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Grisart, B. , Farnir, F. , Karim, L. , Cambisano, N. , Kim, J.J. , Kvasz, A. et al. (2004) Genetic and functional confirmation of the causality of the DGAT1 K232A quantitative trait nucleotide in affecting milk yield and composition. Proceedings of the National Academy of Sciences of the United States of America, 101, 2398–2403. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Gruber, A.J. & Zavolan, M. (2019) Alternative cleavage and polyadenylation in health and disease. Nature Reviews. Genetics, 20, 599–614. [DOI] [PubMed] [Google Scholar]
  16. Heinz, S. , Benner, C. , Spann, N. , Bertolino, E. , Lin, Y.C. , Laslo, P. et al. (2010) Simple combinations of lineage‐determining transcription factors prime‐regulatory elements required for macrophage and B cell identities. Molecular Cell, 38, 576–589. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Jiang, J. , Cole, J.B. , Freebern, E. , Da, Y. , VanRaden, P.M. & Ma, L. (2019) Functional annotation and Bayesian fine‐mapping reveals candidate genes for important agronomic traits in Holstein bulls. Communications Biology, 2, 212. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Jiménez‐González, V. , Ogalla‐García, E. , García‐Quintanilla, M. & García‐Quintanilla, A. (2019) Deciphering GRINA/Lifeguard1: nuclear location, Ca2+ homeostasis and vesicle transport. International Journal of Molecular Sciences, 20(16), 4005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Li, H. & Durbin, R. (2009) Fast and accurate short read alignment with burrows‐wheeler transform. Bioinformatics, 25, 1754–1760. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Li, L. , Huang, K.L. , Gao, Y.P. , Cui, Y. , Wang, G. , Elrod, N.D. et al. (2021) An atlas of alternative polyadenylation quantitative trait loci contributing to complex trait and disease heritability. Nature Genetics, 53, 994. [DOI] [PubMed] [Google Scholar]
  21. Li, L. , Li, Y. , Zou, X. , Peng, F. , Cui, Y. , Wagner, E.J. et al. (2022) Population‐scale genetic control of alternative polyadenylation and its association with human diseases. Quantitative Biology, 10, 44–54. [Google Scholar]
  22. Liao, Y. , Smyth, G.K. & Shi, W. (2014) featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics, 30, 923–930. [DOI] [PubMed] [Google Scholar]
  23. Liu, D. , Xu, Z. , Zhao, W. , Wang, S. , Li, T. , Zhu, K. et al. (2022) Genetic parameters and genome‐wide association for milk production traits and somatic cell score in different lactation stages of Shanghai Holstein population. Frontiers in Genetics, 13, 940650. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Liu, S. , Gao, Y. , Canela‐Xandri, O. , Wang, S. , Yu, Y. , Cai, W. et al. (2022) A multi‐tissue atlas of regulatory variants in cattle. Nature Genetics, 54, 1438–1447. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Park, S.Y. , Kim, S.H. , Yoon, H.K. , Yim, C.H. & Lim, S.K. (2016) The role of nuclear factor‐E2‐related factor 1 in the oxidative stress response in MC3T3‐E1 osteoblastic cells. Endocrinol Metab, 31, 336–342. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Pertea, M. , Kim, D. , Pertea, G.M. , Leek, J.T. & Salzberg, S.L. (2016) Transcript‐level expression analysis of RNA‐seq experiments with HISAT, StringTie and ballgown. Nature Protocols, 11, 1650–1667. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Robinson, M.D. , McCarthy, D.J. & Smyth, G.K. (2010) edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics, 26, 139–140. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Rosen, B.D. , Bickhart, D.M. , Schnabel, R.D. , Koren, S. , Elsik, C.G. , Tseng, E. et al. (2020) De novo assembly of the cattle reference genome with single‐molecule sequencing. GigaScience, 9(3), giaa021. Available from: 10.1093/gigascience/giaa021 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Seif, F. , Khoshmirsafa, M. , Aazami, H. , Mohsenzadegan, M. , Sedighi, G. & Bahar, M. (2017) The role of JAK‐STAT signaling pathway and its regulators in the fate of T helper cells. Cell Communication and Signaling, 15, 23. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Shabalin, A.A. (2012) Matrix eQTL: ultra fast eQTL analysis via large matrix operations. Bioinformatics, 28, 1353–1358. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Sherry, S.T. , Ward, M.H. , Kholodov, M. , Baker, J. , Phan, L. , Smigielski, E.M. et al. (2001) dbSNP: the NCBI database of genetic variation. Nucleic Acids Research, 29, 308–311. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Shulman, E.D. & Elkon, R. (2020) Systematic identification of functional SNPs interrupting 3'UTR polyadenylation signals. PLoS Genetics, 16, e1008977. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Stegle, O. , Parts, L. , Durbin, R. & Winn, J. (2010) A Bayesian framework to account for complex non‐genetic factors in gene expression levels greatly increases power in eQTL studies. PLoS Computational Biology, 6, e1000770. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Strucken, E.M. , Laurenson, Y.C. & Brockmann, G.A. (2015) Go with the flow‐biology and genetics of the lactation cycle. Frontiers in Genetics, 6, 118. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Szklarczyk, D. , Kirsch, R. , Koutrouli, M. , Nastou, K. , Mehryary, F. , Hachilif, R. et al. (2023) The STRING database in 2023: protein‐protein association networks and functional enrichment analyses for any sequenced genome of interest. Nucleic Acids Research, 51, D638–D646. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Tang, Y. , Zhang, J. , Li, W. , Liu, X. , Chen, S. , Mi, S. et al. (2024) Identification and characterization of whole blood gene expression and splicing quantitative trait loci during early to mid‐lactation of dairy cattle. BMC Genomics, 25(1), 445. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Tian, B. & Manley, J.L. (2017) Alternative polyadenylation of mRNA precursors. Nature Reviews. Molecular Cell Biology, 18, 18–30. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Turner, S.D. (2018) qqman: an R package for visualizing GWAS results using QQ and manhattan plots. Journal of Open Source Software, 3, 731. [Google Scholar]
  39. Walker, R.L. , Ramaswami, G. , Hartl, C. , Mancuso, N. , Gandal, M.J. , de la Torre‐Ubieta, L. et al. (2020) Genetic control of expression and splicing in developing human brain informs disease mechanisms (vol 179, pg 750.e1, 2019). Cell, 181, 484. [DOI] [PubMed] [Google Scholar]
  40. Wang, H. , Yu, L. , Cheng, L. & Guo, Z. (2023) The roles of lncRNAs in Th17‐associated diseases, with special focus on JAK/STAT signaling pathway. Clinical and Experimental Medicine, 23, 3349–3359. [DOI] [PubMed] [Google Scholar]
  41. Wathes, D.C. , Becker, F. , Buggiotti, L. , Crowe, M.A. , Ferris, C. , Foldager, L. et al. (2021) Associations between circulating IGF‐1 concentrations, disease status and the leukocyte transcriptome in early lactation dairy cows. Ruminants, 1, 147–177. [Google Scholar]
  42. Xiang, R. , Hayes, B.J. , Vander Jagt, C.J. , MacLeod, I.M. , Khansefid, M. , Bowman, P.J. et al. (2018) Genome variants associated with RNA splicing variations in bovine are extensively shared between tissues. BMC Genomics, 19, 521. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Xue, X. , Hu, H. , Zhang, J. , Ma, Y. , Han, L. , Hao, F. et al. (2022) Estimation of genetic parameters for conformation traits and milk production traits in Chinese Holsteins. Animals: An Open Access Journal from MDPI, 13(1), 100. Available from: 10.3390/ani13010100 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Yang, J. , Lee, S.H. , Goddard, M.E. & Visscher, P.M. (2011) GCTA: a tool for genome‐wide complex trait analysis. American Journal of Human Genetics, 88(1), 76–82. Available from: 10.1016/j.ajhg.2010.11.011 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Zheng, X.W. , Levine, D. , Shen, J. , Gogarten, S.M. , Laurie, C. & Weir, B.S. (2012) A high‐performance computing toolset for relatedness and principal component analysis of SNP data. Bioinformatics, 28, 3326–3328. [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

Appendix S1.

AGE-57-0-s001.docx (24.6KB, docx)

Data Availability Statement

The raw RNA sequence data reported in this paper have been deposited in the Genome Sequence Archive in National Genomics Data Center, China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences (GSA: CRA012735). The public GWAS summary statistics from Figshare (https://figshare.com/s/ea726fa95a5bac158ac1). The SNP BeadChip data for the current study are available from the corresponding author on reasonable request.

Articles from Animal Genetics are provided here courtesy of Wiley

RESOURCES