Mutation columns in metadata

[theosanderson]

A common pattern has been to have columns with comma-separated lists of mutations from reference in "metadata" TSVs, e.g. COG-UK and I think Nextstrain did this. This could be a nice feature for LAPIS. It is particularly useful in the early days of an outbreak.

I'm mostly aware of it for SC2 (so non-segmented) where nucleotide mutations were A123T,G1234A and amino acid mutations were S:P123Q,N:Q235N etc. One could imagine using a pattern as with the amino acid mutations in the case of segmented nucleotide mutations.

I think this would already be useful as a single column per mutation type, which I think is mostly what I've seen so far. (But one could also imagine a more complex thing where you could define what gene/segment you were interested in or something: sorry I don't have detailed thoughts there yet.)