afrubin
commented
7 months ago This seems related to this existing issue about adding descriptive optional metadata for additional columns: https://github.com/VariantEffect/mavedb-api/issues/38
We should also consider adding some kind of controlled vocabulary or set of recommended names for commonly-included columns, so that we don't end up with a bunch of different CRISPR guide or variance estimate columns with similar but distinct names.
We should add support for non-numeric columns. There are two current use cases for this.
The first is allowing relatively short strings that describe variant classes from datasets (e.g. "wild-type like", "possibly damaging"). This would require minimal validation.
The second is allowing users to upload CRISPR guide sequences as part of the score table, accompanying the associated variant. We would like to be able to support this and also enable special validation downstream.
A distinct but related feature that is needed for this is to enable column-wise validation for columns with specific names. We already do this for the hgvs variant columns, but making that logic generic would have additional benefits for future extensions.