Stream: genomics
Topic: variant start-end
Patrick Werner (Jun 06 2019 at 12:39):
looking on mapping start and end of variants to our IG some difficulties were discovered:
1.: 81254-5 Genomic allele start-end is defined as "is the first genomic position in the reference allele" but our dataType is Range, so we should probably switch this to Integer?
Patrick Werner (Jun 06 2019 at 12:41):
2. for CNV/SVs we need start and end, currently we only have inner and outer start-end, but not "exact start-end" or simpler "start-end" do you agree that we need a component here?
Jamie Jones (Jun 06 2019 at 15:28):
For 1, I think the idea is the whole component is a range, so start and end are represented inside it as both ints.
For 2, we're focused on the fuzzy-boundary use case. There should be textual guidance for exact boundary but I don't see it currently... My recollection was the decision was you would set inner and outer the same.
Patrick Werner (Jun 06 2019 at 15:48):
For 1, I think the idea is the whole component is a range, so start and end are represented inside it as both ints.
but this component only has a position, no range.
@2
if i look at the ncbi definitions: https://www.ncbi.nlm.nih.gov/dbvar/content/overview/ i come to the conclusion that semantically we can't set inner and outer to the same value to express exact breakpoints:
"The breakpoints lie inside of the defined region." vs "The breakpoints lie outside of the defined region" if we set them the same the position would lay in and outside at the same time. But maybe this isn't a real problem.
But i can also imagine a use case where i want to transport the actual or estimated breakpoints and the inner outer ones as well.
After reading:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3531204/
I would say we would need a component for actual breakpoint complemented by an confidence intervall extension?
Jamie Jones (Jun 06 2019 at 16:10):
If we know the exact endpoints, can we not use 81254-5? I was never sure if there was a need to distinguish genomic vs structural at that point. Confidence intervals are another matter...
Last updated: Apr 12 2022 at 19:14 UTC
