biomarkerTMLE_exposure {biotmle}R Documentation

TMLE procedure using ATE for Biomarker Identication from Exposure

Description

This function performs influence curve-based estimation of the effect of an exposure on biological expression values associated with a given biomarker, controlling for a user-specified set of baseline covariates.

Usage

biomarkerTMLE_exposure(Y, W, A, a, subj_ids = NULL, g_lib, Q_lib,
  cv_folds = 5, ...)

Arguments

Y

A numeric vector of expression values for a single biomarker.

W

A Matrix of numeric values corresponding to baseline covariates to be marginalized over in the estimation process.

A

A numeric vector of discretized exposure vector (e.g., from a design matrix whose effect on expression values is of interest.

a

The numeric value indicating levels of A above against which comparisons are to be made.

subj_ids

A numeric vector of subject IDs to be passed directly to tmle when there are repeated measures; measurements on the same subject should have the exact same numerical identifier. The correction performed utilizes a more conservative estimator of the variance based on the efficient influence function.

g_lib

A character vector identifying the library of learning algorithms to be used in fitting the propensity score P[A = 1 | W].

Q_lib

A character vector identifying the library of learning algorithms to be used in fitting the outcome regression E[Y | A, W].

cv_folds

A numeric scalar indicating how many folds to use in performing targeted minimum loss estimation. Cross-validated estimates are more robust, allowing relaxing of theoretical conditions and construction of conservative variance estimates.

...

Additional arguments passed to tmle in fitting the targeted minimum loss estimator of the average treatment effect.

Value

TMLE-based estimate of the relationship between biomarker expression and changes in an exposure variable, computed iteratively and saved in the tmleOut slot in a biotmle object.


[Package biotmle version 1.10.0 Index]