An atlas of chicken long non-coding RNAs gathering multiple sources : gene models and expression across more than twenty tissues

International audienceLong non-coding RNAs (lncRNAs) appear as key regulators of gene expression through numerous mechanisms. Yet, particularly in domesticated species like chicken, their annotation is incomplete and their expression in different tissues poorly characterized. Here, we used data from 4 public databases (Ensembl, NONCODE, NCBI, ALDB), a recent lncRNA catalogue, and genes models from 364 RNA-seq samples from our lab to generate an extensive annotation of the chicken genome aggregating most of the currently known lncRNAs, plus robust models of newly discovered lncRNA from our dataset. Our extended lncRNAs annotation comprises 30084 lncRNA loci versus 4640 in the Ensembl reference. These lncRNAs follow the same trends as those in the reference for their genomic localization, structure and expression. To further characterize these lncRNAs, we studied their expression in over 20 chicken tissues with minimum 6 biological replicates. This allowed us to explore expression variability of both lncRNAs and protein-coding genes (mRNAs) across tissues, and across replicates in each tissue. Across these chicken tissues, we found interesting differences in the tissue-specificity patterns between lncRNAs and mRNAs. We characterized the expressions of both types of genes in relation with (i) their tissue-specificity and/or (ii) their genomic localization relative to each other (convergent, divergent, antisense, distant or proximal) in order to highlight some lncRNA–mRNA couples with a role in the functions or the maintenance of a given organ in chicken. Finally, we started a conservation study of interesting lncRNA–mRNA couples across species. Project funded by ANR-13-ADAP and H2020-Feed-a-Gene projects