GE_bias_normal_squaredmis: GE_bias_normal_squaredmis.R

Description Usage Arguments Value Examples

View source: R/GE_bias_normal_squaredmis.R


A function to calculate the bias in testing for GxE interaction, making many more assumptions than GE_bias(). The additional assumptions are added to simplify the process of calculating/estimating many higher order moments which the user may not be familiar with.
The following assumptions are made:
(1) All fitted covariates besides G (that is, E, all Z, and all W) have a marginal standard normal distribution with mean 0 and variance 1. This corresponds to the case of the researcher standardizing all of their fitted covariates.
(2) All G are generated by means of thresholding two independent normal RVs and are centered to have mean 0.
(3) The joint distributions of E, Z, W, and the thresholded variables underlying G can be described by a multivariate normal distribution.
(4) The misspecification is of the form f(E)=h(E)=E^2, and M_j=W_j^2 for all j. In particular, W always has the same length as M here.


GE_bias_normal_squaredmis(beta_list, rho_list, prob_G, cov_Z = NULL,
  cov_W = NULL, corr_G = NULL)



A list of the effect sizes in the true model. Use the order beta_0, beta_G, beta_E, beta_I, beta_Z, beta_M. If G or Z or M is a vector, then beta_G/beta_Z/beta_M should be vectors. If Z and/or M/W do not exist in your model, then set beta_Z and/or beta_M = 0.


A list of expectations (which happen to be covariances if all covariates are centered at 0) in the order specified by GE_enumerate_inputs(). If Z and/or M/W do not exist in your model, then treat them as constants 0. For example, if Z doesn't exist and W includes 2 covariates, then set cov(EZ) = 0 and cov(ZW) = (0,0). If describing expectations relating two vectors, i.e. Z includes two covariates and W includes three covariates, sort by the first term and then the second. Thus in the example, the first three terms of cov(ZW) are cov(Z_1,W_1),cov(Z_1,W_2), cov(Z_1,W_3), and the last three terms are cov(Z_3,W_1), cov(Z_3,W_2), cov(Z_3,W_3).


Probability that each allele is equal to 1. Since each SNP has two alleles, the expectation of G is 2*prob_G. Should be a d*1 vector.


Should be a matrix equal to cov(Z) or NULL if no Z.


Should be a matrix equal to cov(W) or NULL if no W.


Should be a matrix giving the *pairwise correlations* between each SNP in the set, or NULL. Must be specified if G is a vector. For example, the [2,3] element of the matrix would be the pairwise correlation between SNP2 and SNP3.


A list with the elements:


The asymptotic values of the fitted coefficients alpha.


The same beta_list that was given as input.


The list of all covariances (both input and calculated) for use with GE_nleqslv() and GE_bias().


List of calculated means for f(E), h(E), Z, M, and W for use with GE_nleqslv() and GE_bias().


List of calculated Higher Order Moments for use with GE_nleqslv() and GE_bias().


GE_bias_normal_squaredmis( beta_list=as.list(runif(n=6, min=0, max=1)),
rho_list=as.list(rep(0.3,6)), cov_Z=1, cov_W=1, prob_G=0.3)

GEint documentation built on Aug. 10, 2017, 1:03 a.m.