KissDE and batch effect

Question

[david.b.rombaut]

Hello,

due to "reasons", I have a somewhat important batch effect in a cohort, that is impacting quite negatively my splicing analysis. Is there any way that this could be "dealt with" in KissDE (or with another way ?), like it can be done with DESeq2 (modeling the effect by including it into the design formula : ~ batch + condition) ?

By the way, the new --type1-only parameter in KisSplice 2.6.2 is quite a nice addition, thanks!

Best,

David.

Answer 1

[audric.cologne]

Hello,

I do not think this is yet possible with kissDE, even though it could be a quick, easy and nice addition to the package! Maybe in the near future...?

Oh, you noticed that option? Cool! Indeed, it can save a lot of time for people only interested in AS events (even though some AS events could be found in the type 4...).

Have you noticed the Shiny Interface that comes with the lastest version of kissDE? We hope that it can help users to quickly navigate and identify interesting entries in kissDE results...

Best, Audric

[david.b.rombaut]

Hello,

then I will wait for a future update !

Although I noticed the new rds file with writeOutputKissDE, I didn't even know there was a Shiny interface. And wow, it looks quite handy, really cool addition, thanks !

Best, David

[david.b.rombaut]

Hello,

I come back to this "old" question : any news on this subject ?

I have a cohort of paired samples, and the "patient" effect is quite heavy. Similarly to a batch effect, in DESeq2 I can use a design like this : ~ patient + treatment, but still not in kissDE.

I've tried other splicing analysis tools that specifically take paired samples into account, but I must say, results are a little underwhelming.

Best, David