Modeling of fusion assay

[ahwagner]

Several members of 12/15 call mentioned importance of indicating whether an assayed fusion is directly detected or inferred from an underlying assay.

Also, a need to describe the underlying rearrangement if available. Also, there needs to be a need for indicating if the breakpoints for that rearrangement are not known.

[ahwagner]

In email, @dsonkin proposes translation of "inferred" information from assayed fusion data model to a standard "not_assayed" term in nomenclature.

Perhaps we can use optional "not_assayed" term for fusion nomenclature which could be used if desired instead of “?”. In such approach “?” use is preserved and option to be more precise is also available. Does this sound like a reasonable approach?

[ahwagner]

Laveniya Satgunaseelan responds with additional perspective:

I wanted to raise a couple of things that I learnt from others’ perspectives:

1. There is international variation in the way FISH is reported in different countries. Given this variation, I agree that the introduction of a new symbol may add confusion as opposed to clarity. To distinguish between assayed and categorical fusions, I think Dmitriy’s suggestion sounds great. The other suggestion I made to Alex after the meeting was to continue with ‘?’. In the setting of an assayed fusion, one would list ‘inferred’ as fusion evidence, with the inclusion of the ECO ID ontology for the assay method. For FISH, ‘inferred’ would hopefully help the reader interpret the ‘?’ as a limitation of the assay. This approach relies on the reader being clued in to the limitations of each assay.
2. Gokce’s point – should we be curating fusions where we don’t know the partner? In my opinion, I think so as so many community labs use inferred methods (e.g. for ALK) , but I think it’s an interesting point.

[ahwagner]

I think there is broad agreement that, for assayed gene fusions, it would be best to continue to use “?” in the nomenclature. We can also provide additional recommendations around language to be used in assay reports, perhaps something to the effect of:

?::ALK(hgnc:427) fusion inferred from ALK break-apart assay evidence (ECO:0001047)

Here the fusion nomenclature remains compact and consistent with other standards, but additional information specific to clinical assay reporting conforms to recommended language and allows us to express other information content from our model.