Gene Symbols matching to many genes - how lax should symbol aliasing be?

We need to match symbols to genes in a specific annotation release (eg RefSeq 109) - and people are very lax on what gene symbols they use.

We currently look up all the aliases when attempting matches, which can lead to some really old genes coming through.

We need to work out:

* How strict should we be with aliases? Should we walk through them?
* Should we allow a manual override for whether a symbol matches a gene?

Previous conversation at:

https://github.com/SACGF/variantgrid_sapath/issues/117#issuecomment-847618335

Start with the previous conversation in variantgrid_sapath issue 117 and then trace the current lookup that checks all aliases. Done means agreeing and documenting how strict alias matching should be, whether aliases are walked, and how any manual override would work for a specific annotation release.

Written by the indexing model from the issue text.