Genome wide association study identifies novel potential candidate genes for bovine milk cholesterol content

Abstract

This study aimed to identify single nucleotide polymorphisms (SNPs) associated with milk cholesterol (CHL) content via a genome wide association study (GWAS). Milk CHL content was determined by gas chromatography and expressed as mg of CHL in 100 g of fat (CHL_fat) or in 100 mg of milk (CHL_milk). GWAS was performed with 1,183 cows and 40,196 SNPs using a univariate linear mixed model. Two and 20 SNPs were significantly associated with CHL_fat and CHL_milk, respectively. The important regions for CHL_fat and CHL_milk were at 41.9 Mb on chromosome (BTA) 17 and 1.6-3.2 Mb on BTA 14, respectively. DGAT1, PTPN1, INSIG1, HEXIM1, SDS, and HTR5A genes, also known to be associated with human plasma CHL phenotypes, were identified as potential candidate genes for bovine milk CHL. Additional new potential candidate genes for milk CHL were RXFP1, FAM198B, TMEM144, CXXC4, MAML2 and CDH13. Enrichment analyses suggested that identified candidate genes participated in cell-cell signaling processes and are key members in tight junction, focal adhesion, Notch signaling and glycerolipid metabolism pathways. Furthermore, identified transcription factors such as PPARD, LXR, and NOTCH1 might be important in the regulation of bovine milk CHL content. The expression of several positional candidate genes (such as DGAT1, INSIG1 and FAM198B) and their correlation with milk CHL content were further confirmed with RNA sequence data from mammary gland tissues. This is the first GWAS on bovine milk CHL. The identified markers and candidate genes need further validation in a larger cohort for use in the selection of cows with desired milk CHL content.

Figure 1

Manhattan plot of genome-wide significant (p < 5E-05) and suggestive (p < 5E-04) SNP associations for milk cholesterol content in Canadian Holstein cows. The most significant SNPs with their corresponding p-values are indicated. CHL_fat: mg of cholesterol in 100 gram of fat, CHL_milk: mg of cholesterol in 100 gram of milk.

Figure 2

Linkage disequilibrium (LD) pattern on a 1.4–3.4 Mb region of BTA 14. LD blocks are marked with triangles; values in boxes are LD (squared correlation coefficient, r²) between SNP pairs; red boxes indicate LOD > 2 and D′ = 1 (LOD is the log of the likelihood odds ratio, a measure of confidence in the value of D′, where D′ is the ratio of the linkage disequilibrium coefficient D to its maximum possible).

Figure 3

Protein-protein interaction network created using the STRING database for CHL_fat positional candidate genes. Network analysis was set at medium confidence (STRING score = 0.4). The line widths represent the level of interactions (wider lines represent stronger evidence of interactions). CHL_fat: mg of cholesterol in 100 gram of fat.

Figure 4

Protein-protein interaction network created using the STRING database for CHL_milk positional candidate genes. Network analysis was set at medium confidence (STRING score = 0.4). Line widths represent the level of interactions (wider lines represent stronger evidence of interactions). CHL_milk: mg of cholesterol in 100 gram of milk.