Closed rick-heig closed 2 years ago
Version :
sav -v
sav v2.0.1
Hi Rick,
I believe this is already fixed with https://github.com/statgen/savvy/commit/a6f39a4884f3c21881a15a21735dee033e2b3d7b. I tested your example with the latest from the master branch and the genotype in question remains diploid after the round trip. Though I can reproduce the issue by using the v2.0.1 release. A new release is well overdue (my apologies).
Side note: you will need to manually set a --sparse-threshold
that is less than 1.0 in order for --pbwt GT
to have any effect. My current recommendation is --sparse-threshold 0.01
. Better documentation of this is definitely needed.
Thank you for the comments, I get the recommendation for the threshold, I was just testing some edge cases with the very small test file and forgot to set a threshold but found this issue, so I shared it. Good to know it is already taken care of and will be fixed in the next release.
Best regards. Rick
Hello.
I have been running a small test file to see if mixed ploidy (1 and 2) and mixed phasing variant lines are supported and found out that there seems to be an edge case where an allele is lost.
Here is the input (test_file.vcf) :
After running :
I get :
The two following lines differ :
The last sample is now haploid instead of diploid on that variant loci, an allele is lost (
0|0
becomes0
).Best, Rick