Transcriptomics

My first academic works took place in the transcriptomics field. I used to be a member of the KisSplice development team, which is dedicated to the de novo discovery of variants (alternative splicing, SNPs, indels...) in short reads from transcriptomic sequencing, and also puts important efforts into proposing tools for the analysis of such variants. I prolonged my work on this topic by switching from short reads to long TGS reads that in my opinion are promising for alternative splicing studies. I currently work towards a pipeline for de novo study using long, spurious reads. This includes techniques to detect and cluster reads coming from a same gene, but also correction taking advantage of this specific kind of data, as state of the art tools rather focus on genomic reads.

However, I advocate that short reads are still of high interest in the transcriptomics context, because of the coverage levels we can actually only reach through them and because they still provide a quality at least an order of magnitude higher, which is necessary for instance when studying functional impact of a variant. I also believe the full potential of such data has not been reached, both in terms of methods to develop (among possibilities mate-pair and paired-end reads being rarely taken into account, to name but two) and in terms of original application cases (such as [3]). It is also very recently that studies started to investigate deeply the fundamental question of the relative contribution of de novo approaches and reference-bases approaches using short reads, as well as assembly versus mapping benefits.