Reannotation of the zebra finch genome reveals undiscovered transcriptional complexity

ABSTRACT The zebra finch has become a paradigmatic model system to mechanistically investigate vocal communication from genes to behavior. Here, we report 1) a comprehensive map of the polyadenylated fraction of the transcriptome of the male zebra finch telencephalon at two ages, during the song learning phase at 50 days and when fully adult at 2 years as well as 2) as an expanded and refined annotation of the zebra finch genome based on our transcriptome data. Using high‐throughput next generation sequencing of paired‐end strand‐ specific cDNA fragments, we detected over 50% of the annotated protein coding transcripts (ENSEMBL v.55) in the telencephalon at both stages. We exploited the information gained from the paired‐end sequence reads and from reads falling onto splice junctions to update the existing annotation (ENS55) with new transcript structures and alternative splicing events. The dataset allowed to refine 2822 of the available gene models by extending them in the 3’‐UTR and to detect 11391 undiscovered transcriptional units in non‐coding regions. Our results illustrate the value of continuously incorporating new transcriptional evidence into existing annotations.