search

timbitz / Whippet.jl

Lightweight and Fast; RNA-seq quantification at the event-level
MIT License
105 stars 21 forks source link

Not showing concordance with other tools, can we manually normalize? #70

Closed seyfim closed 6 years ago

seyfim commented 6 years ago

Hello,

My current research involves implementing alternative splicing software to identify candidate genes for downstream analysis. Currently, I have implemented MAJIQ, rMATS and your tool. While we see 40-60% concordance between MAJIQ and rMATS, we are seeing < 6% with those tools and Whippet. We also had another tool that was transcript based, Kallisto, and in order to see concordance with other transcript based tools I had to implement upper quartile normalization on the Raw Read counts before continuing DE analysis.

Will normalizing the Raw_Reads column in the .gene.tpm files (generated from Whippet Quant) affect/influence the delta psi calculation?

timbitz commented 6 years ago

Hi @seyfim, reporting of "issues" for this repository are reserved for actual software bugs, and are not the intended forum for questions on program usage, command line options, and output interpretation. For this we ask that you use the gitter chat, or you can e-mail me directly tim.sterne.weiler@utoronto.ca and I would be more than happy to try and help.

To quickly answer your question though: No, nothing in the .tpm file will affect deltaPsi calculations. However, altering the significance and filtering criteria for deltaPsi could definitely make a difference.. As could ensuring that you are intersecting the coordinates from Whippet with other programs correctly (Whippet "nodes" vs. "exons" are often not directly comparable, please see the README.md for info on this).

mirimia commented 6 years ago

Hi Tim and @seyfim,

Sorry to comment on this closed issue, but I’m quite interested on this topic too. If you email about it, please CC me if you don’t mind (mirimia@gmail.com).

Thanks