FaSTLMM.flmm
— Functionflmm: fit linear mixed model using grid of values
y: 2-d array of (rotated) phenotypes X: 2-d array of (rotated) covariates lambda: 1-d array of eigenvalues ngrid: number of grid values to consider
FaSTLMM.flmm
— Functionflmm: fit linear mixed model
y: 2-d array of (rotated) phenotypes X: 2-d array of (rotated) covariates lambda: 1-d array of eigenvalues reml: boolean indicating ML or REML estimation
FaSTLMM.rotateData
— MethodrotateData: Rotates data with respect to the kinship matrix
y = phenotype matrix X = predictor matrix K = kinship matrix, expected to be symmetric and positive definite
FaSTLMM.wls
— Functionwls: Weighted least squares estimation
y = outcome, matrix X = predictors, matrix w = weights (positive, inversely proportional to variance), one-dim vector
The variance estimate is maximum likelihood
FaSTLMM.rss
— Methodrss: residual sum of squares
y = outcome, matrix X = predictors, matrix
Calculates the residual sum of squares using a QR decomposition. The outcome matrix can be multivariate in which case the function returns the residual sum of squares of each column. The return values is a vector of length equal to the number of columns of y.