used for 10x data - Githubissues

GoekeLab / proActiv

Estimation of Promoter Activity from RNA-Seq data

https://goekelab.github.io/proActiv/

Other

48 stars 14 forks source link

used for 10x data #40

Closed houruiyan closed 2 years ago

houruiyan commented 2 years ago

Dear Dr. I found a ppt from you in 2021 BioC. You display the proActiv applying in smart-seq2 data. Any suggestion to deal with smart-seq2 data? In addition, I want to use your package in 10x. I used the pseudo bulk. However, even if I relex the threshold , it still cannot detect any alternative promoter? Any suggestion? Thank you very much!

houruiyan commented 2 years ago

Here, how do you did the preprocessing (i.e. filtering)? Thank you!

jleechung commented 2 years ago

Hi @houruiyan

For filtering we adopted a standard approach:

Filtered out cells expressing less than a certain number of genes
Filtered out genes expressed in less than a certain number of cells

You can try to tweak these thresholds and see how it affects the results.

We've only tried proActiv on full-length single cell sequencing protocols (like SMART-seq). We surmise it might be difficult to get results from 10X 3' data as there are fewer junction reads toward the 5' end.