Add conventions for creating a variant name with punctuation and capitalization

[kkrysiak]

1) For umbrella variant fusion names where the fusion partner is unknown/unspecified we prefer the style "GENE Fusions" or "GENE FUSIONS" (e.g., ALK Fusions). I see all combinations of "GENE Fusion", "GENE Fusions", "GENE FUSION", "FUSION", etc. Example in docs is "ALK Fusions" but I think "ALK FUSIONS" is preferred? There isn't really a leader in current curation: https://civicdb.org/search/variants/6ee62717-61f4-465a-b783-b3f2fc0ade52

2) For specific fusion names we prefer the style "GENEA-GENEB" and not "GENEA-GENEB Fusion" or "GENEA-GENEB FUSION". Again, we have a mix

Similar confusion for rsIDs (rs12345 vs RS12345)

Is Loss-of-function or Loss of function or title case?

[kkrysiak]

Delins c. p. g.

[malachig]

We should also have comprehensive examples for simpler variants in the table: https://civic.readthedocs.io/en/latest/model/variants/name.html

Right now we do not have a variant name example for:

• A simple SNV
• A small insertion
• A small deletion
• A germline variant with a "c." designation in parentheses

Also I wonder if we should revisit the idea of using rs ids as variant names?

[malachig]

Also we don't have an example of how we name a frameshift variant. Currently I see a mix of the following styles:

• F137fs
• I109FS
• G104fs*158
• P146Gfs*14
• R167fsX4

We could adopt the simplest and most easily readable style? F137fs. The HGVS expressions for the variant will show the correct longer style.

However, there could be multiple distinct frameshifts that start at that amino acid position. This would perhaps argue for the most precise style: P146Gfs*14. Its also the most consistent with HGVS.

[malachig]

Show how/when to use "and" and "or" in variant names.

• ERBB2 S310F/Y -> ERBB2 S310F or S310Y
• EML4-ALK C1156Y and L1196M
• MTOR E2014K and E2419K

Note that this is a temporary measure until genotypes / variant sets are supported in CIViC

[kkrysiak]

Clarify to only use an rsID if it occurs outside of a coding / transcribed region. Using c. or p. is preferred.

Avoid more ambiguous names like 3'UTR polymorphism (https://civicdb.org/events/genes/2596/summary/variants/433/summary/evidence/1031/summary#evidence)

[kkrysiak]

• DYPD homozygotes are not mentioned.
• TP53 truncating mutation (bucket variant type variant)
• Variants which exclude a region (ABL1 Non-P-loop mutation)
• Splicing alteration (c.340+1G>A)
• Partial deletion of 0.7 Kb - the size of the deletion can be included in the evidence item and not the variant name

[malachig]

Splicing alterations we could just shorten the name to include the c. notation only. Rather than saying "Splicing alteration".

e.g. "Splicing alteration (c.340+1G>A)" would become "c.340+1G>A".

Also make it clear that we do not want any use of the "IVS" style of splicing variant names.

[malachig]

State a convention or guidance on when to use "LOSS" vs "DELETION" vs "UNDEREXPRESSION".

[kkrysiak]

Variant names for internal duplications when we have gone beyond FLT3-ITD. https://civicdb.org/events/genes/29/summary/variants/67/summary#variant

[kkrysiak]

Exon 14 skipping would be another example to handle.

[kkrysiak]

We should make sure we also provide guidance for variant type consistency, particularly for splice altering variants rather that don't make a change to the DNA.

[lsheta]

Working table with variant name conventions: