Mining RNA Sequencing Data for Evidence of Allele Specific Expression in Cattle

Allele specific expression (ASE) is an epigenetic event that results in random, maternal or paternal preferential expression of an allele at a heterozygous locus. DNA methylation and imprinting are believed to be the primary cause.  Detection of ASE is important because it commonly occurs in genes functioning in growth and survival and likely regulates gene expression networks. The objective of this study is to identify ASE from RNAseq data.  Sequence data from three tissues was aligned to the UMD3.1 bovine reference with bwa.  GATK was used for variant detection, tracking base quality and allele counts.  Variants were compared across three tissues within an individual to identify tissue-specific variants.  Tissue-specific variants observed as such across all individuals (N=8) were selected as candidate ASE variants.  More than 55,000 variants within >9000 transcripts were identified (Q>=20).   Since a proportion of these variants were likely observed due to alternative splicing, candidate ASE variants were filtered such that only transcripts containing greater than three candidate ASE variants were retained in an effort to enrich for ASE.  More than 4400 variants met these criteria.   Multiple genes known to exhibit ASE were identified (e.g. IGF2, IGF2R, IGF1, PEG3, PEG10).  Interestingly, transcripts containing candidate ASE variants were significantly enriched for biological processes related to the trait of interest in the RNAseq experiment, fatty acid metabolism (i.e. insulin signaling, adipokine signaling, FDR < 0.05).  Additional methods are being tested to determine how to best detect ASE and distinguish it from alternative splicing.  This project is supported by the USDA-AFRI-2011-67012-30692.