It is a fascinating and clever idea get microbial genomes of your interest separated from a complex community! It basically works by treating the same environmental or community sample in different ways (treatments) and then extract the DNA to further sequence them. Due to the treatment, there should be differences in the microbial abundance in these smaples by which you can separate your interesting microbial bins on a coverage bi-plot (for example).

You can take a quick look at how I did Differential Coverage Binning from my metagenomic samples from my GitHub page Binning