Knowledge-Based Reconstruction of mRNA Transcripts with Short Sequencing Reads for Transcriptome Research

While most transcriptome analyses in high-throughput clinical studies focus on gene level expression, the existence of alternative isoforms of gene transcripts is a major source of the diversity in the biological functionalities of the human genome. It is, therefore, essential to annotate isoforms of gene transcripts for genome-wide transcriptome studies. Recently developed mRNA sequencing technology presents an unprecedented opportunity to discover new forms of transcripts, and at the same time brings bioinformatic challenges due to its short read length and incomplete coverage for the transcripts. In this work, we proposed a computational approach to reconstruct new mRNA transcripts from short sequencing reads with reference information of known transcripts in existing databases. The prior knowledge helped to define exon boundaries and fill in the transcript regions not covered by sequencing data. This approach was demonstrated using a deep sequencing data set of human muscle tissue with transcript annotations in RefSeq as prior knowledge. We identified 2,973 junctions, 7,471 exons, and 7,571 transcripts not previously annotated in RefSeq. 73% of these new transcripts found supports from UCSC Known Genes, Ensembl or EST transcript annotations. In addition, the reconstructed transcripts were much longer than those from de novo approaches that assume no prior knowledge. These previously un-annotated transcripts can be integrated with known transcript annotations to improve both the design of microarrays and the follow-up analyses of isoform expression. The overall results demonstrated that incorporating transcript annotations from genomic databases significantly helps the reconstruction of novel transcripts from short sequencing reads for transcriptome research.