Identification of long non-coding RNA in the horse transcriptome

Background: Efforts to resolve the transcribed sequences in the equine genome have focused on protein-coding RNA. The transcription of the intergenic regions, although detected via total RNA sequencing (RNA-seq), has yet to be characterized in the horse. The most recent equine transcriptome based on RNA-seq from several tissues was a prime opportunity to obtain a concurrent long non-coding RNA (lncRNA) database. Results: This lncRNA database has a breadth of eight tissues and a depth of over 20 million reads for select tissues, providing the deepest and most expansive equine lncRNA database. Utilizing the intergenic reads and three categories of novel genes from a previously published equine transcriptome pipeline, we better describe these groups by annotating the lncRNA candidates. These lncRNA candidates were filtered using an approach adapted from human lncRNA annotation, which removes transcripts based on size, expression, protein-coding capability and distance to the start or stop of annotated protein-coding transcripts. Conclusion: Our equine lncRNA database has 20,800 transcripts that demonstrate characteristics unique to lncRNA including low expression, low exon diversity and low levels of sequence conservation. These candidate lncRNA will serve as a baseline lncRNA annotation and begin to describe the RNA-seq reads assigned to the intergenic space in the horse.