sub-labels for annotation rows

JAL-2226 (2.11.2) subtly changed the way that 'reference annotations' for a sequence are discovered and added to the alignment, so that several annotations with the same label and calcId could be shown for a sequence.

Shading by sequence associated annotation was (slightly) broken by this - one solution is to support aggregate/averaging over all rows with a particular label/calcId (JAL-3986).

The display of sequence-associated annotation was also made less useful. Multiple annotation rows of the same name are not visually distinguished, and adding the possibility of more annotations exacerbates the problem. JAL-4091 was proposed in 2.11.3 to address this - labelling each associated row with its reference sequence ID. This works when row labels are distinct, but when multiple rows (e.g. 'Secondary Structure') are present it is also necessary to distinguish each row.

Showing the description for each row with the same label & reference sequence is one solution. This would reflect the same logic as 'Add Reference Annotation'. Descriptions include the PDB ID or structure data filename. This is somewhat arbitrary, however, and may not work as expected since the description can include html and can be quite lengthy.

Instead, annotation rows derived from Sequence related data such as 3D structures could receive a sub-label that describes what the row data was derived from.

That would mean annotation rows have the following relational model:
[SeqReference] Label Sub-Label CalcId