FMT tutorial: improve longitudinal and other questions

[nbokulich]

Improvement Description
Elaborate on some questions to suggest different time point comparisons, and/or add notes about issues with low sample size making these tests not entirely adequate.

Current Behavior
Sounds like some of the questions in the FMT tutorial do not pan out too well with the 1% subsample dataset (10% subsample may be a different story) (see forum xref). In particular, low sample size (N = 1 or 2) with the 1% makes the results of some questions confusing.

Proposed Behavior
Note that low sample sizes with 1% (and possibly 10%) subsample datasets may be too low for some tests, particularly paired comparisons between specific time points.
Propose testing a range of different time points (see suggestions in forum xref)

References
forum xref