How to normalise the metabolic gene counts obtained from direct metagenome assembly

[srisvs33]

Thanks for developing this wonderful package to streamline metabolic functional annotations. I have a general query. I have performed shallow metagenome sequencing for 20 seawater samples. The sequence output total reads were varied among each sample, resulting in metagenome assembly sizes ranging from 100 mb to 600 mb in size. Since it’s a shallow metagenome, I am planning to directly annotate the metagenome assembly using DRAM for various metabolic gene functions (gene counts). However, I have a query on how to normalise the gene counts, since the assembly size varies across different samples. What is the best possible method/approach to do this? Could you please provide some insights?
Many thanks.
Venkat