When I tried running additive logistic model using the following command, I always get one predictor for the genotype (A1). But since there are 3 values (0,1,2) for the counts of minor alleles, there should be two predictors associated with SNP for the output model right? How do I make plink do that?
Another question is that I always get NAs for all the output statistics. Got really frustrated on this.
I saw in many tutorial they have the flags --genotypic used with --logistic. However when I am trying to run --logistic --genotypic or with the two flags switched, --genotypic always gets deprecated. Anyone knows why this happened?
The code I was basically using is: plink --bfile filename.bed --logistic --make-pheno covfile.txt 1
0 answers
No answers yet.
Log in to answer this question.
-
PLINK error doing a GWAS
written by david.badenas.01 •
-
recessive model in saige
written by ccagg
-
Adding additional covariates to plink logistic regression model on HLA returns p-vals etc as NA
written by 4galaxy77
-
How to fix phenotype information when converting from VCF to PLINK
written by Diego •
-
Error with CONVERTF mergeit program
written by harmadikemil •
-
Dominance effect SNP estimate
written by joao_gabrielzp •
-
PLINK and probABEL results differ
written by lucia •
-
How to include interaction in logistic model using plink
written by lucia •
-
Problem with logistic regression in Plink
written by C-man •
-
How To Caculate The Or Value Of Additive Model With Plink
written by bluegenechip •
Don't worry about the "deprecated" warning message; that's just informing you that the preferred syntax is now "--logistic genotypic" instead of "--logistic --genotypic". But the latter still works.
However, if all your results are NAs, something else is wrong. The first thing I'd check is the phenotype values; are they being set properly? (--make-just-fam provides one way to dump the phenotype values plink is using.) If they're all missing, or identical, that explains the all-NA result.