Posted: Thu May 21, 2015 4:57 pm
Can the GRAMMAR-gamma method (Svishcheva et al., 2012) be used with imputed genotypes (dosages)? If so, would this be run under the GenABEL package? Is it also possible to obtain robust standard errors with the GRAMMAR-gamma method? When using the GRAMMAR-gamma method for models with a gene-environment interaction, can you obtain estimates of the covariance between the SNP and the interaction?