A 660-Kb deletion with antagonistic effects on fertility and milk production segregates at high frequency in Nordic Red cattle: additional evidence for the common occurrence of balancing selection in livestock

Abstract

In dairy cattle, the widespread use of artificial insemination has resulted in increased selection intensity, which has led to spectacular increase in productivity. However, cow fertility has concomitantly severely declined. It is generally assumed that this reduction is primarily due to the negative energy balance of high-producing cows at the peak of lactation. We herein describe the fine-mapping of a major fertility QTL in Nordic Red cattle, and identify a 660-kb deletion encompassing four genes as the causative variant. We show that the deletion is a recessive embryonically lethal mutation. This probably results from the loss of RNASEH2B, which is known to cause embryonic death in mice. Despite its dramatic effect on fertility, 13%, 23% and 32% of the animals carry the deletion in Danish, Swedish and Finnish Red Cattle, respectively. To explain this, we searched for favorable effects on other traits and found that the deletion has strong positive effects on milk yield. This study demonstrates that embryonic lethal mutations account for a non-negligible fraction of the decline in fertility of domestic cattle, and that associated positive effects on milk yield may account for part of the negative genetic correlation. Our study adds to the evidence that structural variants contribute to animal phenotypic variation, and that balancing selection might be more common in livestock species than previously appreciated.

Figure 1. Association scan of the bovine genome for QTL influencing fertility.

A. Genome-wide Manhattan plot obtained by across-breed single-point analysis for an index combining fertility traits (number of inseminations in heifers and cows (AISH and AISC), interval between calving and first insemination (ICF) and interval between first and last insemination in heifers and cows (IFLH and IFLC)). B. Chromosome-specific (BTA12) plots corresponding to within-breed single-point analyses for the same fertility index (red: Finnish Ayrshire, orange: Swedish red, blue: Danish red, green: Jersey and gray: Holstein cattle). The horizontal lines mark the genome-wide significance threshold.

Figure 2. Characterization of the BTA12 deletion.

A. Characterization of the deletion using SNPs from the BovineHD genotyping array (Illumina, San Diego, CA): (a) average signal intensity (LRR) and (b) mean heterozygosity per marker for carriers (red) and non-carriers (gray). B. Characterization of the deletion at the sequence level: (a) Integrative Genomics Viewer (IGV – http://www.broadinstitute.org/igv/) screen capture of NGS data featuring the 3 Mb region encompassing the deletion with, from top to bottom, tracks of depth coverage for a wild-type (gray) and a carrier (red) animal, a track of RefSeq gene annotation and a RNA-seq junctions output track obtained with TopHat ; (b) zoom of the gene content within the 660 kb deletion, where transcriptional units are depicted as transcription-oriented hatched red arrows above RNA-seq data; (c) sequence trace of the 281 bp amplicon bridging the breakpoint (BP); (d) alignment of deleted sequence (Del BP) with wild-type proximal (Ref pBP) and distal (Ref dBP) BP sequences highlighting the “C” insertion (red) and the 2 bp microhomology (underlined); gray boxes correspond to wild-type sequences within the deletion, directly adjacent to the respective BP boundaries.

Figure 3. Fine-mapping of the fertility QTL on BTA12.

A. Haplotyped-based QTL fine-mapping results on BTA12 for the trait “interval from first to last insemination of cows (IFLC)” in Finnish Ayrshire (red) and Swedish Red (gray) cattle. The x-axis represents the physical position on BTA12 and the y-axis the p-value of the likelihood ratio test. Full and dashed lines represent QTL mapping with and without correction for haplotype B28 tagging the deletion. The vertical dotted lines indicate the position of the deletion whereas the horizontal line marks the chromosome-wide significance threshold. B. Effect (in days) and frequency of the 40 ancestral haplotypes in Finnish Ayrshire (red) and Swedish Red (gray) cattle.

Figure 4. Increased reproductive failure rate in mating types sorted according the genotype of sire and maternal grand-sire for the 660 Kb deletion (I for non-carrier x non-carrier, II for non-carrier x carrier, III for carrier x non-carrier and IV for carrier x carrier matings defined as in the main text), at 35, 56, 100 and 150 days post-insemination.

The default level of reproductive failure was set as the average level observed for mating types I and II (as no offspring homozygous for the deletion can be produced in these matings). The excess failure rates that were observed in the different mating types and at the different time-points are shown by the red bars. The excess failure rates that are expected assuming no development of homozygous conceptuses (calculated as described in the text and assuming a frequency of 0.128 for the deletion) are shown by the gray bars.

Figure 5. Detection of a QTL influencing milk production and composition on BTA12 in Nordic Red breeds.

A. Haplotyped-based QTL mapping on BTA12 for milk (gray), protein (blue) and fat (orange) yield in Nordic Red cattle. The x-axis represents the physical position on BTA12 and the y-axis the p-value of the likelihood ratio test. Full and dashed lines represent QTL mapping with and without correction for haplotype B28 in strong association with the deletion. The vertical dotted lines indicate the position of the deletion whereas the horizontal line marks the chromosome-wide significance threshold. B. Effect and frequency of the 40 ancestral haplotypes on milk (gray), protein (blue) and fat (orange) yield (in Kg). Haplotype B28 tags the deletion.