Hi all, I performed a GWAS using the Rvtest software (model: logistic regression with genotype and a number of covariates as predictors). Next I want to extract the significant SNPs and perform further sensitivity analyses. I used vcftools to extract the SNP (coded as 0, 1 or 2 for AA, AB, BB) from vcf file and loaded this into R with the vcfR package. As a first step I tried to reproduce the model and test statistic performed by Rvtest. However, I get different beta, SE, and p-value. (First, the effect was also in the other direction, this turned out to be due to coding of cases and controls as 1/0 instead of 2/1 as in Rvtest.) My guess is that the single-variant score test that Rvtest performs is different from the beta (=logit) in R output. Do you know how to reproduce the single variant score test in a logistic regression model in R? Thanks in advance, Vincent
