Dear all,
I am running a Gene X environment interaction analysis, where I am testing if the interaction of the phenotype with an environmental variables that shows any significant association.
For this I am implementing the Robust Standard error option implemented in probabel ---- that actually accounts for the variability seen in environmental variable. In general, the literature suggests to use Robust SE when considering interaction analysis in probabel.
But as of now we are testing whether the robust option is making any sense in our dataset, so we ran interaction (association) analysis using probabel with both robust SE and normal SE.
The lambdas were then calculated using R --- using Robust SE (lambda = 1.06) and using Normal SE (lambda=0.80)
My question here is whether the robust option is correcting the extreme deflation seen in using the normal SE. And can we rely on using the robust option in this case.
Any advise will be great.
