Clinical-Genomics / genmod

Annotate models of genetic inheritance patterns in variant files (vcf files)
http://moonso.github.io/genmod/
MIT License
74 stars 17 forks source link

GENMOD output gives pattern of inheritance for familes with no genotype data #55

Closed tinuthomas closed 8 years ago

tinuthomas commented 8 years ago

I am using GENMOD to get my VCF annotated with patterns of inheritance

GENMOD version genmod version: 3.4.8

Commands used genmod annotate TEST.vcf --annotate_regions >TEST_genmod_step1.vcf genmod models TEST_genmod_step1.vcf --family_file TEST.ped > TEST_genmod_step2.vcf

I am getting genetic models assigned for families with no genotype data or just with homozygous reference genotype The VCF had members from many families and accordingly made a PED file with multiple families and that was used to run GENMOD

Family Sample GT A60 S714 ./. A60 S715 ./. A60 S716 ./.

A60 PEDfile

FamilyID SampleID Father Mother Sex Phenotype

A60 S714 S715 S716 1 2 A60 S715 0 0 1 1 A60 S716 0 0 2 1

Result from GENMOD output GeneticModels=A60:AR_hom_dn|AD_dn|AR_hom|AD

Why is this happening ? Am I doing something wrong ?

moonso commented 8 years ago

Hi,

the philosophy is that it will not exclude a genetic model if there is no proof against it. If you want the opposite, that is you want proof for a genetic model to get it annotated, use the flag --strict.

Good luck!

Måns

tinuthomas commented 8 years ago

Thanks, I tried with '--strict'. Now the number of families listed in the GeneticModels reduced.

For this particular example variant, I have the AC=2 and the genotype happens to be in 2 samples B39A and B39C (0/1 for both) and B39B has the genotype '0/0'

FamilyID SampleID Father Mother Sex Phenotype

B39 B39A B39B B39C 2 2 B39 B39B 0 0 1 1 B39 B39C 0 0 2 1 L008 L008B 0 0 1 1 L008 L008C 0 0 2 1 L006 L006B 0 0 1 1 L006 L006C 0 0 2 1

GENMOD output results GeneticModels=L008:AR_comp|AR_hom|AD,L006:AR_comp|AR_hom|AD

The families L008 and L006 listed have 2 members each and both of them are unaffected and all 4 samples have the genotype '0/0'. These members are unrelated. I used '--strict' this time, why are the results so ?

moonso commented 8 years ago

Hmm it is a bit hard since I can't see all data. The families that you refer to, L008 and L006, is not in the pedigree file that you show in the example. Perhaps you could show more data in the issue or send me some files with dropbox or email?

tinuthomas commented 8 years ago

Also there is another family with 4 members with a kid, dad, mom and brother, with only the kid affected with the disease. All 4 of them have the genotype '0/0'. Again '--strict' option was used. How can I interpret the results ?

SampleID GT (Family - 51) 516(kid) 0/0 517(dad) 0/0 518(mom) 0/0 519(brother) 0/0

GENMOD output 51:AR_hom_dn|AD_dn|AR_comp_dn

moonso commented 8 years ago

Again it is hard for me to tell what's going on without the data. If it is like you say something is wrong... But we use this tool alot and I do not recognice these prolems