is there a way to give protein sites non-sequential numbering?

### Context In Nextstrain trees, is there a way to give protein sites non-sequential numbering for things like labeling branches and coloring by protein site? Right now I understand how you can pass a GFF to say define HA1 and HA2 to start at some position within the full coding sequence. But people often use H3 numbering for other subtypes (like H7 and H5). Here the numbers of sites in the H7 or H5 HA are not always strictly sequential: if H7 has a deletion relative to H3 then the H3 numbering skips a site, and if H7 has an indel then there is a duplicate number (eg, site 133a). PDB files are often numbered this way. Is there any established way to do this within Nextstrain trees? ### Description As an example, we currently use [H3 numbering](https://dms-vep.org/Flu_H5_American-Wigeon_South-Carolina_2021-H5N1_DMS/numbering.html) for mutations in a H5 HA. This mapping of numbering is shown [here](https://github.com/dms-vep/Flu_H5_American-Wigeon_South-Carolina_2021-H5N1_DMS/blob/main/data/site_numbering_map.csv) (see *reference_site* and *sequential_site* columns). ### Possible solution Perhaps one possible solution is to define a very complex GFF that defines a non-sequential feature, although this could not capture all forms of indels between the numbering schemes. Not sure if @lmoncla @kistlerk @Haddox @jordan-ort have dealt with this?