NCM 2021: Bambu — generating context-aware transcriptomes with Oxford Nanopore long-reads
- Home
- NCM 2021: Bambu — generating context-aware transcriptomes with Oxford Nanopore long-reads
Andre stated that the great thing about the field of transcriptomics is how innovative and dynamic it has been: from using microarrays a few years ago, to now ‘here we are at the Nanopore Community Meeting, talking about the cutting edge of RNA sequencing technology’. To help address some key problems involved in transcript discovery and the development and use of transcriptome reference annotations, Andre’s team have developed Bambu, which is a two-module workflow that couples transcript discovery with transcript quantification. Andre’s talk focused on the transcript discovery module, the basis of which is an R package, and he demonstrated how Bambu could successfully classify and identify novel transcripts. Explaining how transcript discovery is strongly affected by sequencing depth, Andre showed how Bambu stabilises sample variance, and compared its performance in terms of precision and sensitivity with other transcript discovery tools (Flair, Talon and Stringtie2). Bambu demonstrated higher sensitivity and precision, and over a greater dynamic range — ‘this shows that Bambu is a very effective tool for transcript discovery, regardless of if you want to be super-sensitive…or if you want to be a lot more conservative and precise’. Explaining why reference annotations are so important, Andre described how they could impact downstream processes, such as quantification. To this end, Andre demonstrated how Bambu’s improved transcript discovery also improved transcript quantification, i.e., estimates of isoform expression levels were found to be more accurate than other approaches; this improvement in quantification accuracy was not just for novel transcripts, but also for those transcripts that were already known and included within the reference. Andre stated that Bambu has been developed and evaluated using direct RNA, direct-cDNA, and PCR-cDNA nanopore sequencing data, and they like to think that it is a ‘very easy to use’ and intuitive tool.