How to performe differential gene expression analysis for a single patient

Question

[nhaus]

Hey everybody,

I think this is a basic question, but somehow I am really confused right now and it would be amazing if somebody could point me the right way.

I was wondering how it is possible to perform DGE for a single cancer patient. I thought, that if you use programs like DESeq2, you need biological replicates to perform the statistical analysis. And if you are only investigating one patient you do not have different replicates. If we take several samples from the same tumor, this would be considered technical replicates right? Is there a way to circumvent this problem, or is it just not possible, to do this for a single individual patient?

Any help is appreciated!

Thank you very much, Nick

Answer 1

[JC]

Several samples of the same tumor/patient are indeed technical replicates.

You still can do DEG analysis with that, however your p-values will be not valid or considered, better is you put a high threshold in fold-change ratio and minimal trusted value.

[nhaus]

Thank you for your answer! Could you maybe elaborate what you mean with minimal trusted value?

I have never heard of it in the context of DGE analysis

[JC]

consider that you have this:

Gene  Test  Control
A         6       1
B         600    100

both cases, you a FC=6X, but the first case is not so trusted because you have fewer counts that validate your numbers, when you do this with replicates, this affects the p-value, but because you don't have replicates and p-value, you need to filter the values to consider only those with a minimal number of reads and increase the confidence.

[nhaus]

That makes a lot of sense, thank you!

Do you know if the lfcshrink function of DESeq2 works without replicates? Because I think this would also help in determining "significantly" differentially expressed genes.