It will be helpful to know what kind of shifts? Do you want to see changes in the structure or species composition or functional shifts of a specific group? Based on that the techniques may be different or require more than one type of analysis.

DGGE will give you an overall picture of the structure of the microbial community without giving you the identity of the individuals present except you excise the bands and sequence. NGS will give you identity of the individuals present. Because DGGE is relatively cheaper, what most people do is to first do a DGGE and if they see certain patterns in the community structure, they will go ahead to do NGS. In our lab, we tend to combine both to get a holistic picture of the community. Check our publications.

Fatma Hassaneen I agree with Israel Ikoyi . you can do DGGE analysis as its cheap. 1st you have to set different time point and during treatment and also before the treatment, so you will have a comparative analysis. if you find the difference then you can go for NGS.

I would reccommed microbial 16s amplicon libraries (for bacteria) and/or 18s for eukaryotes (ngs) these techniques are much cheaper than before and give you some idea of what might be causing the change (i.e. loads of e.coli probably = anthropgenic contam) for functional shifts shotgun metagenomics may be your best bet but it is pricy, complicated and requires huge amounts of computing power to assemble contigs (without assembly it is difficult to accurately determine function).

DGGE works for determining if there are shifts (i.e yes or no) but tells you little else and is a pretty outdated method if you plan yo publish.

I support Eric opinion. 16S and some conservative functional genes sequencing will give you much more information comparing to DGGE.