We should have a way to add notes to a dataset that are conveyed to users

### Context

We are now digging into requests for datasets from users. Some require additional engineering work to import, but others are unprocessed because of fundamental limitations of the data or mistakes in metadata.

### Problem or idea

We should have some way to note these and convey them to our users. Ideally we'd have a box that would show up in place of the request dataset button that would have a text field that we could put messages. For example, for the dataset in #2128 it would be:
"We have looked into importing this dataset. We are unable to import it because the underlying data are incomplete. The investigator appears to have uploaded the sample indices but not the full reads."

### Solution or next step

I think this would require adding a field to the dataset model. I guess this could pose an issue if we wanted to make notes only about certain, unprocessed samples. On the other hand, cases like this (or the incorrectly labeled drop-seq experiment that @cansavvy found), the whole experiment should really be suspect.

Start by locating the dataset model and the request dataset button described in the issue, then trace how dataset details are stored and presented to users. Decide whether notes apply to whole datasets or individual samples; done means a stored note can replace the request button and display the provided explanation to users.

Written by the indexing model from the issue text.