Identification of Tissue-Specific Long Non-Coding RNAs in Three Livestock Species

Long non-coding RNAs (ncRNAs) have been found in eukaryotic organisms to be up to four times more prevalent than coding transcripts and are critical for the regulation of gene expression. As part of the Functional Annotation of Animal Genomes (FAANG) pilot project, we have identified tissue-specific long ncRNAs across eight tissues in chicken, pig and cattle. Using RNA-seq data from two biological replicates per species, transcripts were assembled with Cufflinks and expression levels were calculated using HTSeq. DESeq 2 was used to determine tissue-specific transcripts, which were then filtered to remove single-exon transcripts, as these are often the result of sequencing noise. Finally, CPC (Coding Potential Calculator) was used to separate potential unidentified coding transcripts from non-coding transcripts. An average of 216 tissue-specific transcripts were identified per tissue in cattle, 221 in pig, and 351 in chicken. Some brain tissues displayed generally higher number of tissue-specific long ncRNAs than others (533 in pig cortex, 274 in cattle cerebellum, and 556 in chicken cerebellum). Lung tissue also has an above average number of tissue-specific long ncRNAs (484, 393, and 450 in chicken, cattle, and pig, respectively), likely due to the heterogeneous nature of the tissue. The lowest numbers were observed in liver and muscle tissues. The tissue-specific expression of nearby protein-coding genes indicate a potential regulatory role for some long ncRNAs. While long ncRNAs are known to be less evolutionarily conserved than protein-coding genes, reciprocal BLAST show some evidence of conserved sequences between species, especially pig and cattle.