Is it appropriate to use markers with expression predicted by one model together with markers predicted by another model in the same clustering?

[denvercal1234GitHub]

Hi there,

Thank you again for the tool.

There are markers that have their expression better imputed by one model over the other models.

What are your thoughts regarding whether using markers that have their expressions imputed by 1 model together with markers that have their expression imputed by another model in the same clustering?

Or we should only stick to 1 model that can impute well more markers?

Thank you for your advice.