Illuminating the dark side of the human transcriptome with long read transcript sequencing

摘要

Long read RNA diversity and splice junction wobble. a RNA samples are typically comprised of a mixture of degraded and immature RNA as well as DNA fragments that can be erroneously identified as novel genes and transcripts. Non-classical RNA here represents lncRNA that do not have a 5′ cap or poly-A tail. b Representation of how RNA sample noise appears when mapped to the genome. c Due to sequencing errors, mapping reads can introduce wobble in determination of splice junctions. Wobble is defined as the difference between exon starts/ends between mapped reads. Wobble walking occurs when 3 or more transcript models have exon starts/ends with each closest pair occuring within the wobble threshold but with the outer pair of exon starts/ends having an in between distance greater than the wobble threshold. TAMA Collapse uses several methods of analyzing wobble to identify true splice junctions