Questions on sample size in MintTea application

[Ivy-ops]

Hi Developer,

Thank you for creating this tool. I have two questions:

(1) What is the recommended sample size for applying MintTea? Minimal sample size for each omic is suggested?

(2) If I have datasets from different omics with varying sample sizes but some overlapping samples, does MintTea support this scenario?

Best regards,
-Ivy

[efratmuller]

Hi @Ivy-ops!

(1) While I cannot provide any strict threshold for a reasonable number of samples, I would avoid using MintTea with less than a few dozens of samples. However, you can try running MintTea on your dataset and if indeed it is too small to capture any meaningful multi-omic module, then MintTea isn't supposed to return anything at all. This is due to the repeated subsampling approach: In a small dataset, you are not likely to get the same modules using different subsets of the data, unless there is a very strong signal.

(2) MintTea unfortunately does not support the scenario you describe. It can only be used with samples that have all omic profiles available, so only the intersection of samples in your case.

I hope this helps,
Efrat