search

monarch-initiative / GENO-ontology

Repository for representing genotypes and their association with phenotypes
18 stars 6 forks source link

germline vs somatic variants #19

Open nlwashington opened 9 years ago

nlwashington commented 9 years ago

we will need to disambiguate between variants (and perhaps any geno partonomy) that are germline vs somatic.

how should these be represented? should there be classes for "germline variant" "somatic variant", or perhaps there is some kind of property about the origin of the variation "germline origin". if it's a "somatic origin", i could see perhaps linking the origin to a tissue or cell type, if known.

the usecase for this is particularly for cancers, where major changes to the genome (chromosomal rearrangements) happen in only a subset of the cells but not the germline. also, it's possible that we might want to reason differently based on the origin, and in interfaces filtering/sorting on the origin is important.

we hadn't really needed this before as with mice and fish they were generally somatic, but in humans we definitely need it.

selewis commented 9 years ago

Seems like a property (relationship type?) might be the best way to go. It's easy to imagine all sorts of ways of partitioning variants. germline vs. somatic point vs. cnv vs. rearrangement impact types etc., etc.

On Fri, May 8, 2015 at 6:29 AM, Nicole Washington notifications@github.com wrote:

we will need to disambiguate between variants (and perhaps any geno partonomy) that are germline vs somatic.

how should these be represented? should there be classes for "germline variant" "somatic variant", or perhaps there is some kind of property about the origin of the variation "germline origin". if it's a "somatic origin", i could see perhaps linking the origin to a tissue or cell type, if known.

the usecase for this is particularly for cancers, where major changes to the genome (chromosomal rearrangements) happen in only a subset of the cells but not the germline. also, it's possible that we might want to reason differently based on the origin, and in interfaces filtering/sorting on the origin is important.

we hadn't really needed this before as with mice and fish they were generally somatic, but in humans we definitely need it.

— Reply to this email directly or view it on GitHub https://github.com/monarch-initiative/GENO-ontology/issues/19.

mbrush commented 9 years ago

SO has a sequence attribute hierarchy describing variant origin.

SO_0001762 (variant origin) ---SO_0001781 (de novo variant) ---SO_0001778 (germline variant) ---SO_0001775 (maternal variant) ---SO_0001776 (paternal variant) ---SO_0001779 (pedigree-specific variant) ---SO_0001780 (population-specific variant) ---SO_0001777 (somatic variant)

I suggest we simple create triples that link variants to these attributes. GENO has a property called has_origin (GENO:0000643) we could use here:

< :variant1 GENO:0000643(has_origin) SO:0001777(somatic_variant)>