Studies with mixed data types

[cleme]

Is your feature request related to a problem? Please describe.
Some studies include more than a single microbiome data type, e.g. 16S and shotgun. It is currently unclear whether those should be stored as different studies or the same.Note this is different to studies in which we have microbiome and some other -omic datatype (metabolomics), because in that case the secondary datatype is considered (somewhat arbitrarily) "metadata".

Describe the solution you'd like
This needs to be discussed carefully. On the one hand, allowing multiple data types within the same study is more natural and might allow for cross-comparison of data types (e.g. correlation analysis btw 16S and shotgun). PICRUSt is, in a sense, already doing this, as predicted functions are included as part of a study that has 16S data. However, this will complicate code base even assuming we can call Ensemble from MMEDS. Separating into different studies (StudyX_16S vs StudyX_shotgun) simplifies things but complicates centralizing analysis.