Identification and Use of Regulatory Variation in Dairy Cattle

There is increasing evidence that genetic variation affecting complex traits lies in regions of the genome regulating gene expression. Expression quantitative trait loci (eQTL) mapping and allele specific expression (ASE) are two ways of finding cis-regulatory genetic variation, that could also affect complex traits. We have used RNA sequence data to investigate the extent of ASE in cattle. We conclude that ASE is pervasive in cattle, is most often tissue specific and that more than half the time displays divergent allele specific expression patterns across tissues. In another study we compared SNP significantly associated with gene expression with those significantly associated with variation in complex traits. Results showed that SNP driving ASE were also often local eQTL implying that they were cis-eQTL. These SNP often affected gene expression in more than one tissue and the allele increasing expression was usually the same. Also, SNP significantly associated with gene expression were more likely to influence complex traits. We have also used RNA sequence data from milk and white blood cells (WBC) from 29 Jerseys and 112 Holsteins with imputed whole genome sequence data to identify eQTL that affect gene expression in lactating cattle. eQTL analysis revealed 132 (milk cells) and 649 (WBC) eGenes (p<1x10-6, false discovery rate of 0.016 and 0.013 respectively). There was a trend for the eQTL to be close to the gene transcription start site (TSS), with 23% of significant eQTL within 50kb of the TSS.

Due to the extent of linkage disequilibrium within Holstein cattle, the eQTL have broad peaks and identification of causative mutations is very difficult. The underlying mutations responsible for the eQTL could be identified more precisely if there was good annotation of the bovine genome for regulatory regions (eg promotors and enhancers). With the ultimate aim of identifying these regions, and as a contribution to the Functional Annotation of Animal Genomes (FAANG), two lactating (5^th and 7^th parity) and pregnant (16 weeks) Holstein females and their unborn foetus’ were euthanized and extensive tissue sampling performed. 40 tissues/cell types were sampled from each cow/foetus pair. RNA was successfully extracted from 36 of these tissue/cell types and stranded libraries have been prepared in duplicate. RNA sequencing is underway aiming for 40 million reads per library. Chromatin Immunopreciptation (ChIP) sequencing is currently being optimised for 8 tissues.

Identification of cis-regulatory variants responsible for phenotypic variation in cattle production traits might be accelerated by the above strategy, leading to more rapid identification of causative mutations affecting complex traits, and more accurate genomic selection.