Permanent fix for duplication of merged SNPs on import

[ljwh2]

Duplicate SNPs cause the data release to break and issues in data presentation, please refer to #1275 for more details. Ala cleaned up all the existing examples in the database but code changes are required to stop it happening in future.

Acceptance criteria for this issue:

• Curators can extract merged SNPs using the merged rsID
• unpublished SNPs do not break the DR
• published SNPs appear only once in the UI & download

[ljwh2]

On further analysis, this may be a problem on Ensembl side, as none of the duplicated SNPs are found in Ensembl (though they are found in dbSNP). This issue needs a bioinformatician to look at it.