Abstract
BACKGROUND: Long non-coding RNAs (lncRNAs) are important component of mammalian genomes, where their numbers are even larger than that of protein-coding genes. For example, human (Homo sapiens) (96,308 vs. 20,376) and mouse (Mus musculus) (87,774 vs. 22,630) have more lncRNA genes than protein-coding genes in the NONCODEv5 database. Recently, mammalian lncRNAs were reported to play critical roles in immune response to influenza A virus infections. Such observation inspired us to identify lncRNAs related to immune response to influenza A virus in duck, which is the most important natural host of influenza A viruses. RESULTS: We explored features of 62,447 lncRNAs from human, mouse, chicken, zebrafish and elegans, and developed a pipeline to identify lncRNAs using the identified features with transcriptomic data. We then collected 151,970 assembled transcripts from RNA-Seq data of 21 individuals from three tissues and annotated 4094 duck lncRNAs. Comparing to duck protein-coding transcripts, we found that 4094 lncRNAs had smaller number of exons (2.4 vs. 10.2) and longer length of transcripts (1903.0 bp vs. 1686.9 bp) on average. Among them, 3586 (87.6%) lncRNAs located in intergenic regions and 619 lncRNAs showed differential expression in ducks infected by H5N1 virus when compared to control individuals. 58 lncRNAs were involved into two co-expressional modules related to anti-influenza A virus immune response. Moreover, we confirmed that eight lncRNAs showed remarkably differential expression both in vivo (duck individuals) and in vitro (duck embryo fibroblast cells, DEF cells) after infected with H5N1 viruses, implying they might play important roles in response to influenza A virus infection. CONCLUSIONS: This study presented an example to annotate lncRNA in new species based on model species using transcriptome data. These data and analysis provide information for duck lncRNAs' function in immune response to influenza A virus.