Title: De novo assembly and identification of unique contigs in the bovine oviduct from animals with high and low circulating estradiol concentrations during timed artificial insemination Authors
Submitted to: Society for the Study of Reproduction Annual Meeting
Publication Type: Abstract Only
Publication Acceptance Date: March 29, 2014
Publication Date: N/A
Technical Abstract: Reproductive efficiency is a large concern for many cattle producers and understanding the mechanisms responsible for biological variation in reproduction is key to improving reproductive efficiency. Timed artificial insemination of beef cows with high circulating estradiol concentrations at time of insemination is known to result in an increase in pregnancy rate of 27%±5% compared to animals with low estradiol at the same time point. To explore the role the oviductal environment plays in this reproductive success, we conducted RNA-seq on DNAse treated RNA isolated from the isthmus and ampullary-isthmic junction collected 24 hr after AI from six animals with differential estradiol concentrations (10.12±0.62 pg/ml vs 5.97±0.62 pg/ml, P=0.0093) at time of fixed-time AI (High-E2 n=3, Low-E2 n=3). Paired end libraries for RNA-sequencing were constructed using the Truseq version 2.0 RNA-seq library prep kits and were sequenced on a Hi-seq 2000. Sequencing reads were concatenated within estradiol classification and tissue. De novo assembly of transcripts was performed using the Trinity software package. Assembly of the transcriptome generated ~1.43 million contigs across classification and tissue types. To estimate contig abundance, the paired-end reads used in the de novo assembly were then mapped to the Trinity generated contigs from the High-E2 and Low-E2 groups within tissue using the RSEM pipeline. Contig abundance is conveyed using the fragments per kilobase of exon model mapped per million mapped reads (FPKM) method. Contig identity was assessed using the BLASTn algorithm with the NCBI Ref Seq mRNA, Bos taurus UMD 3.1 cDNA, and the corresponding contrast group within tissue as the search databases. To identify novel contigs, the E-value for the BLASTn search was set at 10, minimizing the probability that these contigs were previously identified sequences (P=0.0001). Across tissue and estradiol classification, we found 298 contigs with FPKM greater than 20 that were not detected in either the NCBI Ref Seq mRNA or the UMD Bos taurus 3.1 cDNA databases. Furthermore, we found that there were nine contigs where expression was greater than 20 FPKM in the High-E2 group, 0 FPKM in the Low-E2 group, and that failed to generate a BLASTn alignment in the Low-E2, UMD Bos taurus 3.1 cDNA, and NCBI mRNA databases. In the Low-E2 group, there were 68 contigs where expression was greater than 20 FPKM, 0 FPKM in the High-E2 group and failed to generate a BLASTn alignment in the High-E2, NCBI mRNA, and UMD Bos taurus 3.1 cDNA databases. To confirm that these contigs were real sequences and not computational artifacts, PCR was conducted on pools of cDNA within tissue-type and estradiol status for a subset of contigs (n=4 per tissue type) using primers specific to the target. Confirmatory PCR detected products for all contigs examined. The contigs described above demonstrate the value of additional analysis when conducting RNA-seq experiments, especially in species where the genome is lacking complete annotation. Finally, these contigs provide unique targets for future studies aimed at understanding the role the oviduct plays during early embryonic development in cows with differential reproductive efficiencies.