search

RitchieLabIGH / IRFinder

MIT License
13 stars 10 forks source link

IRFinder Diff #25

Closed rezarahman12 closed 1 year ago

rezarahman12 commented 1 year ago

Dear IRFinder-S developer team, Thank you for building this nice tool. I followed the instructions for singularity installation. I'm using our inhouse RNA-seq data to identify differential introns. IRFinder worked well for all the steps except when I'm trying to run the differential IR analysis, I'm getting Rscript not found. Below is my code-

./IRFinder Diff -g:TTX ./TTX1/IRFinder-IR-nondir.txt -g:TTX ./TTX2/IRFinder-IR-nondir.txt -g:TTX ./TTX3/IRFinder-IR-nondir.txt -g:CTR ./Ctr1/IRFinder-IR-nondir.txt -g:CTR ./Ctr2/IRFinder-IR-nondir.txt -g:CTR ./Ctr3/IRFinder-IR-nondir.txt -ir 0.05 -o results.dir Rscript not found. I'll highly appreciate how can I run the IRFinder Diff. I'll look forward to hearing at your earliest convenience. Best wishes Reza The University of Queensland, Australia

CloXD commented 1 year ago

Hello Reza, First of all, thank you for using IRFinder-S. To help you, I would need to know what operative system you are using and what image, if the one generated using singularity build or the one you can download from here https://github.com/RitchieLabIGH/IRFinder/releases/tag/v2.0.1 . If there is some conflict with the environmental variables, you could try:

singularity exec -e ./IRFinder Diff ...

Furthermore, there could be an error in R due to the loading of local libraries. You can avoid this by setting the home to an empty directory: 

mkdir ${HOME}/tmp_home/
singularity exec -e -H ${HOME}/tmp_home/ ./IRFinder Diff ....

I hope that one of those solutions will help. Cheers, Claudio

rezarahman12 commented 1 year ago

Hi Claudio, Thank you so much for your quick response. Yea, your suggestions helped. I'm using the HPC cluster Bunya to run IRFinder. However, I had just need to modify a little bit singularity exec -e ./IRFinder IRFinder Diff ... Just one more clarification- the results files created are given in the attached screenshot. Is that okay? or any other plot should be generated as well? Thank you again and appreciate your time and help. Best wishes, Reza image

CloXD commented 1 year ago

Hello Reza, I'm glad it worked. In case there are IR events with FDR < 0.05, a pdf file is created with one boxplot for each event. Cheers, Claudio

rezarahman12 commented 1 year ago

Hi Claudio Thank you so much for your response. Is it possible to generate box plot of IR events based on p-value not FDR since IRFinder in the case of my data did not give any IR events passed FDR <0.05?

Thanks a lot again. Best regards Reza

CloXD commented 1 year ago

Dear Reza, It's not possible using only the IRFinder method, but you can easily generate them using R or Python. You can retrieve the IR ratios for each sample from the file all_results_DESeq2.tsv or from the original IRFinder-IR-nondir.txt files. Cheers, Claudio

rezarahman12 commented 1 year ago

Thanks a lot, Claudio for your suggestions.