I have been doing single cell analysis based on seurat and I realized after each clustering and reclustering when I run SCTransform, in my code there is vars.to.regress = percent.mt and I was running this everytime. I removed MT- contaminated cells before.

My clustering seems fine though.

Do you think this is very critical thing and I have to analyze everything from scratch?


I think it would be a good idea to at least look at the distribution of mitochondrial gene fraction in your cells. If they mostly have low fractions (e.g. less than 15%), then I wouldn't expect the clustering to be influenced too much by them.

With SCTransform regressing on the percent.mt, I think it's okay to do post-transform filtering, but it'd be a good idea to exclude any cells that have high mitochondrial fraction prior to doing clustering (unless, for whatever reason, those are the cells you're interested in).


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.