latent variables

[Flu09]

Hello, i am new to the tool. I am familiar with seurat and limma package but as I understand it is available through muscat as well.
I have gathered samples from different studies (2 different regions affected by the same disease) . Few samples are from the same individual (2 different regions from same individuals) while others are different individuals for each region.

I want to look at different things first disease vs control, disease_genotype_region vs control_genotype_region

I want to ask if I need to add the donor as a latent variable? I am just afraid since it is not confounding factor for some of the samples, it would affect the model negatively.

should I include donor or just include sex, age, etc.. Another question if it I need to include batch or number of genes or UMI counts since they come from different studies so the number of RNA moleculas and genes are different and the batch is different as well?

Thank you.