Feed efficiency in dairy sheep: An insight from the milk transcriptome

Abstract

Introduction: As higher feed efficiency in dairy ruminants means a higher capability to transform feed nutrients into milk and milk components, differences in feed efficiency are expected to be partly linked to changes in the physiology of the mammary glands. Therefore, this study aimed to determine the biological functions and key regulatory genes associated with feed efficiency in dairy sheep using the milk somatic cell transcriptome.

Material and methods: RNA-Seq data from high (H-FE, n = 8) and low (L-FE, n = 8) feed efficiency ewes were compared through differential expression analysis (DEA) and sparse Partial Least Square-Discriminant analysis (sPLS-DA).

Results: In the DEA, 79 genes were identified as differentially expressed between both conditions, while the sPLS-DA identified 261 predictive genes [variable importance in projection (VIP) > 2] that discriminated H-FE and L-FE sheep.

Discussion: The DEA between sheep with divergent feed efficiency allowed the identification of genes associated with the immune system and stress in L-FE animals. In addition, the sPLS-DA approach revealed the importance of genes involved in cell division (e.g., KIF4A and PRC1) and cellular lipid metabolic process (e.g., LPL, SCD, GPAM, and ACOX3) for the H-FE sheep in the lactating mammary gland transcriptome. A set of discriminant genes, commonly identified by the two statistical approaches, was also detected, including some involved in cell proliferation (e.g., SESN2, KIF20A, or TOP2A) or encoding heat-shock proteins (HSPB1). These results provide novel insights into the biological basis of feed efficiency in dairy sheep, highlighting the informative potential of the mammary gland transcriptome as a target tissue and revealing the usefulness of combining univariate and multivariate analysis approaches to elucidate the molecular mechanisms controlling complex traits.

Keywords: RNA-Seq; dairy sheep; feed efficiency; mammary gland; sPLS-DA.

Figure 1

Functional enrichment results from the differential gene expression analysis between high (H-FE) and low (L-FE) feed efficiency animals. In the GOCircle plot, the significant GO terms enriched (FDR < 0.05) after a reduction of the terms with a gene overlap >80% are represented. The outer circle shows a scatter plot for each GO term of the logFC of the genes clustered in the term. The blue circles are genes downregulated in H-FE, while red circles are upregulated genes in H-FE sheep. The inner circle shows a bar plot representing the z-score for each GO term. The red bar means that the GO term is upregulated for H-FE, while the blue bar indicates the GO term is upregulated for L-FE.

Figure 2

Results from the sparse Partial Least Square-Discriminant analysis (sPLS-DA). (A) Sample prediction area plot from the sPLS-DA model applied on the RNA-Seq data set from high (H-FE; orange triangles) and low (L-FE; blue circles) samples using as the distance for prediction “maximum distance”. (B) Loading plot of the top 20 discriminating genes on the first component between high and low feed efficiency animals, colors indicate the group in which the mean expression is maximal for each gene (H-FE: orange and L-FE: blue). (C) GOCircle plot showing the significant GO terms and pathways enriched (FDR < 0.05) after a reduction of the terms with a gene overlap >80% are represented. The outer circle shows a scatter plot for each term of the logFC of the genes clustered in the term. The blue circles are genes downregulated in H-FE, while the red circles are upregulated genes in H-FE sheep. The inner circle shows a bar plot representing the z-score for each term. The red color means that the GO term is upregulated for the H-FE group; the red color intensity is associated with the value of the z-score.