Within the genome, we have chromosome plus position, which are independent fields. From a human perspective, sorting on "position" most logically means the combination of chromosome + position, where chromosome establishes a sort order. To facilitate this, we already compute a "genomePosition" field, which creates a single numeric series to order positions in the "human correct" order.
When the user tries to sort the variant search results on position, I think we should silently use genomicPosition as the actual sort field. Otherwise sorting on position gives weird results like:
Chrom1 1
Chrom2 1
Chrom3 1
....
Chrom1 2
Chrom2 2
Chrom3 2
We could either handle this client-side, or look for this string server-side. I cant think of a reason either is better or worse. I wonder if the chromosome/contig field should have the same behavior.
Within the genome, we have chromosome plus position, which are independent fields. From a human perspective, sorting on "position" most logically means the combination of chromosome + position, where chromosome establishes a sort order. To facilitate this, we already compute a "genomePosition" field, which creates a single numeric series to order positions in the "human correct" order.
When the user tries to sort the variant search results on position, I think we should silently use genomicPosition as the actual sort field. Otherwise sorting on position gives weird results like:
We could either handle this client-side, or look for this string server-side. I cant think of a reason either is better or worse. I wonder if the chromosome/contig field should have the same behavior.