Targeted RNA sequencing (CaptureSeq) uses oligonucleotide probes to capture RNAs for sequencing, providing enriched read coverage, accurate measurement of gene expression, and quantitative expression data. We applied CaptureSeq to refine transcript annotations in the current murine GRCm38 assembly. More than 23,000 regions corresponding to putative or annotated long noncoding RNAs (lncRNAs) and 154,281 known splicing junction sites were selected for targeted sequencing across five mouse tissues and three brain subregions. The results illustrate that the mouse transcriptome is considerably more complex than previously thought. We assemble more complete transcript isoforms than GENCODE, expand transcript boundaries, and connect interspersed islands of mapped reads. We describe a novel filtering pipeline that identifies previously unannotated but high-quality transcript isoforms. In this set, 911 GENCODE neighboring genes are condensed into 400 expanded gene models. Additionally, 594 GENCODE lncRNAs acquire an open reading frame (ORF) when their structure is extended with CaptureSeq. Finally, we validate our observations using current FANTOM and Mouse ENCODE resources.

The authors acknowledge the following funding sources: an Australian National Health and Medical Research Council (NHMRC) Australia Fellowship (631668 to J.S.M. and 631542 to M.E.D.); an NHMRC Early Career Fellowship (APP1072662 to M.B.C.); an EMBO Long Term Fellowship (ALTF 864-2013 to M.B.C.); an Australian National Health and Medical Research Council (NHMRC) Project Grant (APP1062106 to T.R.M.) and Career Development Fellowship (APP1062470 to T.R.M); and an EMBL Interdisciplinary Postdoc (EIPOD) under Marie Curie Actions (COFUND) (to G.B.).