Contamination found in sequencing data of an isolated cell structure, what can I do?

Question

[marina.wakid]

Hi all,

I isolated particular cell types (cell types that comprise a functional structure that can be isolated as a structure) from brain tissue (brain samples belong to two groups). This was more of an enrichment than purification and after sequencing RNA from those structures, I see some contamination from cell types I don't want via bioinformatic deconvolution. The contamination is not huge but because the contamination in itself was not picked up consistently, some genes from contaminating cell types are creeping up in DGE analysis. Is there any kosher way to correct for/account for/circumvent/remove this contamination pre- or post- DGE analysis? Anything to do at all?

Thanks!

[GenoMax]

I see some contamination from cell types I don't want via bioinformatic deconvolution.

If you isolated the cells by following a careful experimental protocol then because "bioinformatics says there is contamination" needs to be taken with a grain (or more) of salt. Trust your experimental expertise.

The contamination is not huge but because the contamination in itself was not picked up consistently, some genes from contaminating cell types are creeping up in DGE analysis

If you think those genes don't make biological sense to address the question you are asking then in your downstream experimental validation experiments you can exclude those.

[marina.wakid]

Thank you for your advice! My main concern comes from working with postmortem tissue which, due to tissue quality, can act inconsistently even with robust experimental protocols, so although I know deconvolution is an estimate, it is both plausible that there is contamination and/or that bioinformatic deconv is a bit off...

There's nothing else that can be done?