Identification and evolutionary analysis of long non-coding RNAs in zebra finch

Abstract

Background: Long non-coding RNAs (lncRNAs) are important in various biological processes, but very few studies on lncRNA have been conducted in birds. To identify IncRNAs expressed during feather development, we analyzed single-stranded RNA-seq (ssRNA-seq) data from the anterior and posterior dorsal regions during zebra finch (Taeniopygia guttata) embryonic development. Using published transcriptomic data, we further analyzed the evolutionary conservation of IncRNAs in birds and amniotes.

Results: A total of 1,081 lncRNAs, including 965 intergenic lncRNAs (lincRNAs), 59 intronic lncRNAs, and 57 antisense lncRNAs (lncNATs), were identified using our newly developed pipeline. These avian IncRNAs share similar characteristics with lncRNAs in mammals, such as shorter transcript length, lower exon number, lower average expression level and less sequence conservation than mRNAs. However, the proportion of lncRNAs overlapping with transposable elements in birds is much lower than that in mammals. We predicted the functions of IncRNAs based on the enriched functions of co-expressed protein-coding genes. Clusters of lncRNAs associated with natal down development were identified. The sequences and expression levels of candidate lncRNAs that shared conserved sequences among birds were validated by qPCR in both zebra finch and chicken. Finally, we identified three highly conserved lncRNAs that may be associated with natal down development.

Conclusions: Our study provides the first systematical identification of avian lncRNAs using ssRNA-seq analysis and offers a resource of embryonically expressed lncRNAs in zebra finch. We also predicted the biological function of identified lncRNAs.

Keywords: Feather development; Zebra finch; lncRNA; ssRNA-seq.

Fig. 1

Overview of the ssRNA-seq transcriptome assembly and lncRNA identification pipeline. a Overview of the ssRNA-seq-based transcript reconstruction pipeline that was used to identify expressed transcripts in sequencing libraries. b The integrative pipeline for the stringent identification of lncRNAs in zebra finch dorsal skins. CPC: coding potential calculator; PLEK: predictor of long non-coding RNAs and messenger RNAs based on an improved k-mer scheme; HMMER (HMMER-3): Profiling protein sequence data using hidden Markov models

Fig. 2

Genomic and expression features of the predicted zebra finch lncRNAs. Genomic features of (a) transcripts length, (b) exon number, and (c) mean phastCons score of zebra finch mRNA, lincRNA, intronic lncRNA, and lncNAT were compared. d The fraction of lncRNAs overlapping with at least one base of a TE (transposable element) in zebra finch, chicken, human, and bovine. Expression features of (e) expression levels and (f) JS scores of zebra finch mRNA, lincRNA, intronic lncRNA, and lncNAT were compared

Fig. 3

Clustering analysis of the expressed genes and the expression heat map. Hierarchical clustering analysis clustered the 13,362 expressed annotated genes and 2,949 unannotated transcripts into 12 clusters (A-L, see Additional files tables for details). The expression levels of each gene are shown as the scaled FPKM values across the six transcriptomes (scaled z-score: red = up-regulation, blue = down-regulation). Three clades (F, G, and L) used for further analysis were labeled in yellow

Fig. 4

Quantitative PCR and sequence conservation analysis of the three identified lncRNAs. a Expression profiles of lncRNA CUFF.19772.1 in E8, E9, and E12 of zebra finch and chicken AD and PD skins. b Expression profiles of lncRNA CUFF.6222.3 in E8, E9, and E12 of zebra finch and chicken AD and PD skins. c Expression profiles of lncRNA CUFF.14902.2 in E8, E9, and E12 of zebra finch and chicken AD and PD skins. Sequence conservations are shown below the quantitative PCR in blue (conservation among amniotes) and green (conservation among birds)