Set up pipeline differential splicing analysis

[hansenp]

This recent paper provides an overview of pipelines that can be used differential splicing analysis:

https://www.biorxiv.org/content/early/2017/06/30/156752

Figure 1 shows an overview of different pipelines. The lower part is for differential splicing analysis.

[hansenp]

It seems that we have data for two different mutants. Given the high number of replicates recommended for RNA-seq (https://www.ncbi.nlm.nih.gov/pubmed/27022035) I would suggest to do a joint analysis for the two mutants.

We should first go for DEXseq first and then possibly also for Cufflinks. If we get moderately consistent results, we are done.

[sheridae]

I think the issue is with the underlying hypothesis here. DexSeq is a good method when dealing with inclusion of specific exons or splice sites. The method focusses on exons, but it is biased against short exons. This includes alternate 5′ or 3′ splice sites, which are defined as separate exons. So it is less efficient at recognising this class of variation. rMATS approach provides a precise description of the type of alternative splicing is especially valuable where the experimental question involves splicing mechanisms. I think this is the question to be interrogated in the case of PPIL1 mutations.

[sheridae]

The problem for me is: DEXSeq maps all of the exons in the canonical transcripts of a gene into a single conflated construct. This makes it difficult to work out what actual types of splicing events are being detected as well as the physical locations of the mis-spliced exons since the analysis in DEXSeq is gene-centric. In contrast, rMATS is exon-centric, so gives this additional information, with more obvious measures of statistical significance (q values and FDR) and effect size (included level difference). I find that this is invaluable data in order to prioritize individual mis-splicing events for wet lab testing using rtPCR, and some validation of splicing in candidate exons is essential. The rMATS output also helps if you want to identify cis-actting splicing factor binding sequences eg by kmer analysis.

[sheridae]

Hi Peter, excel spreadsheet from the Americans RNA-seq samples.xlsx

[hansenp]

I think there DEXSeq performs the differential analysis on the count data. Before this, scripts are applied that take into account splicing and different isoforms (http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0136653).

But maybe rMATS is a good alternative. Hopefully it is not too complicated to use. At least it was cited 177 times.

[sheridae]

I would still be concerned it would miss the alt 5' and 3' splice sites at the very least.

Best Wishes Eamonn Sheridan Professor of Clinical Genetics Leeds Institute for Molecular Medicine St James's Hospital Leeds LS9 7TF

++113 343 8504

---

From: Peter Hansen notifications@github.com Sent: Wednesday, March 28, 2018 7:01 PM To: hansenp/Ppil1 Cc: Eamonn Sheridan; Comment Subject: Re: [hansenp/Ppil1] Set up pipeline differential splicing analysis (#1)

I think there DEXSeq performs the differential analysis on the count data. Before this, scripts are applied that take into account splicing and different isoforms (http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0136653).

[http://journals.plos.org/plosone/article/figure/image?id=10.1371/journal.pone.0136653.g003&size=inline]http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0136653

RNA-Seq Analysis of Differential Splice Junction Usage and ...http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0136653 journals.plos.org Alternative splicing is an important biological process in the generation of multiple functional transcripts from the same genomic sequences. Differential analysis of ...

But maybe rMATS is a good alternative. Hopefully it is not too complicated to use. At least it was cited 177 times.

— You are receiving this because you commented. Reply to this email directly, view it on GitHubhttps://github.com/hansenp/Ppil1/issues/1#issuecomment-376960662, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AcaeNQTF9WQ08IKrdy8_YK1HeV_Q-JHtks5ti8IFgaJpZM4S2zTX.