driving factors of interest

[cuttlefishh]

It would be good to have a list of driving factors we expect to be important in differentiating samples, based on previous work (e.g. Lozupone and Knight's finding about salinity). We can test for these directly. Note: we can also see what falls out from machine learning, etc. A few to start with:

• salinity (saline vs. non-saline)
• life style (host-associated vs. free-living)
• env_matter