How MG-rasst deal with sequencing depth difference when calculating p-value

Question

[wintergb]

I have a question about MG-rast p-value calculation regarding the sample sequencing depth difference. Say Group A has 10 million reads per sample and Group B has 1 million reads per sample, Group A would have more abundance than Group B due to the sequencing depth difference and the p value between Group A and Group B will be significant as well. I read the FAQ and the manual but didn't find any information about how MG-rast deal with the sequencing depth difference when calculate the p-value, or maybe I miss it. Thanks!