Description
We're currently not storing anything about contigs except the ID and length. Unfortunately the contig header meta information seems pretty loosely constrained, and possibly the only option that we have to keep full information about contigs from the header is to add a catch-all contig_metadata
(or something) array that stores this stuff.
e.g. the VCF spec gives an example of
##contig=<ID=20,length=62435964,assembly=B36,md5=f126cdf8a6e0c7f379d618ff66beb2da,species="Homo sapiens">
So here, the contig_metadata field would consist of assembly=B36,md5=f126cdf8a6e0c7f379d618ff66beb2da,species="Homo sapiens"
I guess it would be more helpful if this was formatted as JSON?
I guess the alternate (probably better) approach would be to suggest that converters MAY attempt to create arrays corresponding to these items (e.g. contig_md5, contig_assembly, contig_species here) on a best-effort basis?