search

nledru / RENIN

Regulatory Network Inference with single cell multiomic data
MIT License
5 stars 2 forks source link

Motif matching algorithm used #1

Open r-trimbour opened 1 year ago

r-trimbour commented 1 year ago

Dear @nledru,

First I would like to thank you for your great work, RENIN seems to be a very interesting tool to predict more accurately CREs !

I'm currenlty looking foward to test it and I was wondering which motif matching algorithm you were using/suggesting to associate the TF-motif pairs from cisBP to the CREs. I didn't find it in the article or the RENIN vignette.

I looked a bit at the raw code, where you're using the object "seurat@assays$peaks@motifs" (modeling.R , line 223), but I didn't find if you were proposing a way to create this motifs-object with RENIN or if I should define it before to start the RENIN pipeline analysis.

Thanks a lot for your time :) Best regards, Rémi

nledru commented 1 year ago

Hi Rémi,

Thanks for the catch, I will add more of a vignette on setting up the dataset. I used Signac's AddMotifs function to do this very quickly, something like this:

devtools::install_github("GreenleafLab/chromVARmotifs") library(chromVARmotifs) data("human_pwms_v2") # filtered human motifs from cisBP motifs <- human_pwms_v2 multi <- AddMotifs(multi, genome = BSgenome.Hsapiens.UCSC.hg38, pfm = motifs, assay = "peaks")

Let me know if this causes any trouble for you, Best, Nicolas

r-trimbour commented 1 year ago

Dear Nicolas,

Thanks a lot for your fast reply and very complete answer ! :)

I think that it could indeed be interesting to add this part to the vignette. I'll let you know if I have any other difficulties.

Best, Rémi