Adjust for latent factors, after rotationn
adjust.latent(
corr.margin,
n,
X.cov,
Gamma,
Sigma,
method = c("rr", "nc", "lqs"),
psi = psi.huber,
nc = NULL,
nc.var.correction = TRUE
)marginal correlations, p*d1 matrix
sample size
estimated second moment of X, d*d matrix
estimated confounding effects, p*r matrix
diagonal of the estimated noise covariance, p*1 vector
adjustment method
derivative of the loss function in robust regression, choices are
psi.huber, psi.bisquareand psi.hampel
position of the negative controls
correct asymptotic variance based on our formula
a list of objects
estimated alpha, r*d1 matrix
estimated beta, p*d1 matrix
estimated row covariance of beta, a length p vector
estimated column covariance of beta, a d1*d1 matrix
The function essentially runs a regression of corr.margin ~ Gamma.
The sample size n is needed to have the right scale.
This function should only be called if you know what you are doing.
Most of the time you want to use the main function cate to adjust for confounders.