search

DavisBrian / seqMeta

Meta-Analysis of Region-Based Tests of Rare DNA Variants
7 stars 3 forks source link

Survival Models must have covariates??? #34

Open DavisBrian opened 6 years ago

DavisBrian commented 6 years ago

This works ####survival data cohort1 <- prepCox(Z=Z1, Surv(time,status)~strata(sex)+bmi, SNPInfo = SNPInfo, data =pheno1)

This does not ####survival data cohort1 <- prepCox(Z=Z1, Surv(time,status)~1, SNPInfo = SNPInfo, data =pheno1)

DavisBrian commented 6 years ago

Seems to be with the handling of strata in the model.

nullmodel$strata <- eval(parse(text=rownames(attr(nullmodel$terms, "factors"))[attr(nullmodel$terms, "specials")$strata]), envir=data) # necessary for stratified analysis - 2014-10-07 - HC

evaluates to eval(parse(text=NULL)) which is waiting for input. Typing NULL causes an error so you don't have to crash R.