nFeature is very high in 10x run data

[bl24]

Hello. I'm new to scRNA and I've been looking at the guided PBMC tutorial and using that as a guide on doing analysis for data I've generated from 10x. I had a question regarding the part of the guide where the data is filtered based on QC metrics:

pbmc <- subset(pbmc, subset = nFeature_RNA > 200 & nFeature_RNA < 2500 & percent.mt < 5)

If I try to apply "nFeature_RNA < 2500", that eliminates almost all of my cells. I was wondering if it's normal for 10x data to have such high nFeature values?

[saketkc]

Do you have a distribution plot for all the 3 features that you can share?

[bl24]

Here's my violin plot for the 3 features.

[saketkc]

Looks fine. The idea of filtration in the tutorial is to remove outliers, so the cutoffs are data dependent. You might just use the mt.percent cutoff here.

[bl24]

Got it. Thanks!