Implement Location class to store symbolic genes and gene regions

[reece]

Split GeneLocation proposal from #181 into separate feature.

[reece]

Use case
Provide a symbolic reference to a gene to be used as (with the role of) a Location. For example, "PTEN deletion" or "VHL, 3 copies".

Consideration should also be given to converting symbolic gene locations onto sequence coordinates, how this might happen, and whether such a transformation is required. (Without this, queries like "identify variation within a gene" would be impossible.)

[reece]

Naming Consideration
Current Location subclasses are locations within that coordinate system. For example, SequenceLocation is a location on a Sequence (as coordinate system), not the location of a sequence.

We have been calling the proposed class a "GeneLocation". Under the current naming pattern, this would imply locations on
Gene. However, it's much more challenging to contemplate the addressable components of a gene.

[ahwagner]

Use cases

Concrete examples of the symbolic reference use case described by @reece above.

Gene-specified fusions

• BCR / ABL1 fusion
• KIAA1549 / BRAF fusion
• ALK fusions
• RET fusions

Gene-specified CNAs

• EGFR amplification
• ERBB2 amplification
• PTEN loss

Gene-specified expression variants

• CCND1 overexpression
• RRM underexpression

Gene region-specified alterations

• EGFR Exon 19 deletion
• KIT Exon 11 mutations
• FLT3 tyrosine kinase domain mutation

Gene wild type variants

• EGFR wild type
• KRAS wild type
• APC wild type

Other categorical gene variants

• FLT3 in-frame ITD
• CEBPA mutations

[sarahhunt]

These seem more like definitions of sets of variants, rather than locations.
We have a similar thing in G2P where we seek HMX1 "loss of function", variants

[github-actions]

This issue was marked stale due to inactivity.