feat: add support for over-writing genotypes for males on x non par #1030
Conversation
bpblanken
changed the title
| @@ -24,5 +27,6 @@ class FeatureFlag: | |||
| EXPECT_TDR_METRICS: bool = EXPECT_TDR_METRICS | |||
| EXPECT_WES_FILTERS: bool = EXPECT_WES_FILTERS | |||
| INCLUDE_PIPELINE_VERSION_IN_PREFIX: bool = INCLUDE_PIPELINE_VERSION_IN_PREFIX | |||
| OVERWRITE_SV_MALE_NON_PAR_CALLS: bool = OVERWRITE_SV_MALE_NON_PAR_CALLS | |||
There was a problem hiding this comment.
My understanding of this logic is it will only apply for internal SV callsets. From notes:
cohort specific of the GREGoR project
There was a problem hiding this comment.
I don't know why we wouldn't apply this to all SVs in seqr, presumably we are doing this because of desired seqr search behavior. I don't think its rreasonable for us to write search queries that are able to handle branching logic for determining what a genotype for a sample is based on whether its internal or external data
| from v03_pipeline.lib.model import ReferenceGenome, Sex | ||
|
|
||
|
|
||
| def overwrite_male_non_par_calls( |
There was a problem hiding this comment.
@matren395 tag as resident actual biologist of this code.
There was a problem hiding this comment.
not clear of the context, but - male X non-par would be regions on ChrX that are (biologically) haploid . reasonably confident these are relevant for a few diseases. It's here where all that cool sex-linked-trait-stuff happens
| non_par_interval = hl.interval( | ||
| par_intervals[0].end, | ||
| par_intervals[1].start, | ||
| ) |
There was a problem hiding this comment.
yep! PARs are at the start and the end of ChrX
There was a problem hiding this comment.
I think theres more than one par interval though?
There was a problem hiding this comment.
I think there's two par intervals at the ends of the chromosome, but only one par region [end par1, start par2]
There was a problem hiding this comment.
yep, there's two intervals - one at the start and one at the end of ChrX. If you take the space between the end of the first and the start of the second (what I believe this is doing) , you get the non-PAR portion of ChrX.
all of my knowledge is from https://en.wikipedia.org/wiki/Pseudoautosomal_region
| mt = mt.annotate_entries( | ||
| GT=hl.if_else( | ||
| ( | ||
| male_sample_ids.contains(mt.s) | ||
| & non_par_interval.overlaps( | ||
| hl.interval( | ||
| mt.start_locus, | ||
| mt.end_locus, | ||
| ), | ||
| ) | ||
| ), | ||
| hl.Call([1, 1], phased=False), | ||
| mt.GT, | ||
| ), | ||
| ) |
There was a problem hiding this comment.
how would this handle SVs overlapping the start of a non-PAR regions, and do we want to be handling them this way ?
There was a problem hiding this comment.
and what do the GTs look like elsewise? are they left haploid ?
There was a problem hiding this comment.
I believe overlaps handles cases where the end of the SV is > the start of the non-Par region.
|
I guess the broader question is what is that code ur implementing in the first place |
Implementation of: