update Uniprot loader to add suffix to canonical protein

[rachellyne]

As suggested by Jerven:

<comment type="alternative products">
<event type="alternative splicing"/>
<isoform>
<id>P05067-1</id>
<name>APP770</name>
<name>PreA4 770</name>
<sequence type="displayed"/>
</isoform>


If there is an altermative product comment take the identifier of the "displayed" item as the identifier of the cannonnical isoform
 else take the recordprimaryAccesion and add '-1'


We would have to think carefully about how to model this - we also do not want to lose the Uniprot primary identifier (without the suffix) as this would affect mapping with other sources. We could have the suffixed version as Primary Accession and the main uniprot identifier as the Uniprot Accession. (as we do for non-canonical entries). I think we on;y load sequence for the canonical protein.