Its a question you should have asked before performing your experiments, scientific experiment usually goes, here is the question , here is the hypothesis and lets do some logical experiments to test this hypothesis.

If you do something and hope the data sticks together and magically gives you interesting results, I am afraid you need a lot of luck for that to happen.

I could have misinterpreted your question however.

I think you should have to follow this link may you find analysis for your data. https://www.biostars.org/p/58279/

I agree with Jay siddharth. However now that you have done it already there are a lot of interpretations you can infer from MG-RAST. Like find important commercial enzymes etc. Besides you can assemble  the metagenomic data and analyse using stand-alone tools to verify your results of MGRAST. You can also perform statistical and diversity estimates on the data depending on the sample you have. There are a lot of options and analysis to be done but it all depends on your objective of study. Hope this helps!!!All  d best

Really Jay has nailed the answer, but an idea for follow up analysis - if you've got some sort of comparative data (control vs treatment, different environments, hosts) you could take a look at the viral population in your samples. Whether they change between groups or can be correlated to shifts in the bacterial community would be interesting to see. 

The same thing can be done with bacteria at the strain level.