Single-cell analysis: Exclude mitochondrial genes / ribosomal genes from highy variable genes prior to dimension reduction

[Wilson Lwtan]

Dear all,

I am dealing with single-cell of cardiomyocytes (abundant with mitochondrial).

I am wondering if anyone has experience seeing less interesting genes in the highly variable gene list ? Eg. mitochondrial genes. 

I wondering if I should remove it (if it has no contribution to the biology)? Just wanna get some opinion.

Thank you.

Wilson 

[BigOmics Analytics]

Hi Wilson. 

I guess this question is best to ask in a forum for scRNAseq, like Seurat. Frequently in the preprocessing, mitochondrial and ribosomal genes are either used for filtering the cells. Then if you are sure mitochondrial and/or ribosomal gene content are not related to the biology, you can regress out the contribution of these genes. I have seen that is often done. This should drastically reduce the variance of the MT genes.

Ivo

BigOmics Team