Deep Analysis of Unmapped Reads From Liver Transcriptome of Nellore Cattle

The bovine reference genome is estimated to be 3000 Mb, organized in 29 pairs of autosomes and two sex chromosomes. It was sequenced from DNA samples of a Bos taurus cow known as Dominette. However, Nellore is the most prevalent breed of beef cattle in Brazil originated from India (Bos indicus). We performed an experiment in Nellore cattle and through hepatic RNA-seq we determined the main biological processes involved in feed efficiency. Although we achieve our objectives, part of the data generated by RNA-seq wasn’t used since the reads didn’t mapped to the reference genome, being excluded of further analyses. Thus, here we aimed to deeply analyze the unmapped reads from the liver transcriptome of Nellore cattle. A mean of 28 million reads was obtained and 16% were not mapped to the reference genome. RNA-seq denovo assembly of unmapped reads resulted in 30,777 transcripts with mean identity to protein sequences of 91%. The unmapped reads were assembled mainly to: human, bovine, mouse, rat and other vertebrate species. Interestingly, we found transcripts from 11 viruses, 6 bacteria and 1 protozoan. A transcript from Shigella flexneri were detected, a cellulosic bacterium found in rumen. Also, we found a transcript from Theileria parva, the causative agent of East Coast fever, a costly disease in Africa affecting bovines. This result probably occurred because of similarity of Babesia bovis and T.parva. As seen, a deep analysis of unmapped reads could also reveal further information that should be considered when analyzing RNA-seq data.