search

huyustats / SCATS

A statistical tool to detect differential alternative splicing events using single-cell RNA-seq
20 stars 3 forks source link

Data sets with batch effect #9

Open hkarakurt opened 2 years ago

hkarakurt commented 2 years ago

Hello, Thank you for this amazing tool. I have a question. I have multiple datasets (all 10X chromium) but their experimental procedures are different so basically there is a batch effect. Is there a specific way to overcome batch effect in datasets in SCATS?

Thank you in advance.

zyh4482 commented 2 years ago

So far there is no method to remove batch effect of splicing events in bulk RNA-seq. Not to mention that in scRNA-seq.

hkarakurt commented 2 years ago

Thank you for your answer. Do you think platform or experimental procedure differences affect the results of DAS as well? I may need to merge 2 sets from 10X and Smart-Seq. I believe this is not a possible option for DAS analysis.

Thank you.

zyh4482 commented 2 years ago

Theoretically, it should have batch effect. There are a few reviews on this topic. The most debated topic is the validity of such analysis because of the relatively low sequencing depth of scRNA-seq. One paper mentioned the question about batch effect. Hence, I can't tell you a conclusive answer because there is no common understanding or standard protocol.You should make your own judgement based on your knowledge or could develop your own tools for this purpose.