Protein of interest

Hi, I am quite confused when I was reading the manual.
I do not know how to get the protein of interest for each sample.
I am working with metagenome data to identify mobile genetic elements (MGEs) in my dataset.

So, to create the protein of interest for each sample, I have to do the assembly first, then do the annotation, and then run Blast against the MGEs database. From this, I have to extract the contigs which were assigned to MGE.

Is this the way to get the protein of interest?
It will be annoying if this is the way!

Or can I run Shortbred and obtain the markers only from the database and then use these markers for quantification?

I have the short-read dataset and the MGEs database.

Can you please help me with understanding this?