search

satijalab / seurat

R toolkit for single cell genomics
http://www.satijalab.org/seurat
Other
2.31k stars 920 forks source link

Normalisation options in Seurat Object #8365

Closed LeaLe88 closed 10 months ago

LeaLe88 commented 10 months ago

Good morning,

I am currently preparing a deconvolution, for which I am using scRNA & bulk RNA Seq data. It is recommended that both should be represented in same normalisation space.

Looking at the normalization.method function in Seurat I see the following options: image

I would like to obtain a length normalised method (like RPKM, FPKM, TPM) - how would that be possible here?

Thank you very much!

longmanz commented 10 months ago

Hi, In Seurat we do not support length-based normalization methods, because length information is not stored. You might need to do it with alternative software + your original scRNA-seq data. Which deconvolution method are you using? Does it explicitly ask you to use the same length-based normalization method to your RNAseq and scRNAseq data?

LeaLe88 commented 7 months ago

Sorry for the late reply, just to close this topic some comments: I am using CIBERSORT. I did not find information in the CIBERSORT support regarding which normalisation method is best for scRNA however I read for bulkRNA seq it is best to use TPM counts and I read somewhere that it is good if both are in the same normalisation space. However in the meantime I found this paper https://genomebiology.biomedcentral.com/articles/10.1186/s13059-023-03016-6/figures/4 It says it doesnt really matter if you use no normalisation, lognorm or TMM counts as long as bulk RNA seq uses TPM counts:

image

So I am still a bit confused whether it should be in the same normalisation space or not. Any further opinions appreciated.