This is a test version of Biostars. For the public version, visit https://www.biostars.org.
plink2 tell "--glm" to use dosage instead of hardcalls

I just upgraded form plink1.9 to plink2, because I read that support for "dosage data" is much better. I imputed my variants with minimac4 and I can successfully converted my vcf to the pfile format using this command:

Options in effect:
  --maf 0.05
  --make-pgen
  --out test
  --pheno  phenotype.txt
  --update-sex gender.txt
  --vcf all.vcf.gz dosage=HDS

In the documentation, I read that plink2 automatically makes hardcalls from the dosage data that I imported.

I want to use "--glm" to predict a quantitative phenotype, and I was wondering, if I have to tell plink2 specifically that it is supposed to use the dosage and not the hardcalls for my alleles.

On a similar note, is it also possible to tell "--glm" to use the hardcalls, even if I imported with the dosage setting?

Any help is much appreciated! Thanks

plink2 gwas imputation

1 answer

I got an answer:

--glm always uses dosages when they're available. If you want to base it on hardcalls instead, you can use "--make-pgen erase-dosage" to generate a dataset with only hardcalls, and then run --glm on that.

Hello nhaus,

I see this answer in the PLINK2 documentation, however, I have not been able to do the latter element and tell PLINK2 to use hardcalls instead of dosages.

I tried:
plink2 --bfile file --make-pgen erase-dosage --out file
and then:
plink2 --pfile file --glm hide-covar --pheno pheno.txt --covar-name PC1, PC2, PC3 --pheno-name y --out glm

but still only obtain one line per variant in the glm.dose.glm.linear output file.

Have you happened to figure this out in the last 8 months? Any help would be appreciated!

Hi, Unfortunately, I cant help you out there. I wanted to make sure that PLINK uses dosages instead of the hardcalls, so I never tried that.. Maybe open another question? The plink2-google group is very active!

Log in to answer this question.