In brianpatrickneal/AMKAT: AMKAT: An Adaptive Multivariate Kernel-based Association Test

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "100%"
)

AMKAT

AMKAT (short for Adaptive Multivariate Kernel-based Association Test) is a flexible, nonparametric test for joint association between a multivariate quantitative response and a high-dimensional set of features. It includes the following notable features:

• A feature selection method for filtering noise variables out of the feature set
• A kernel-based test statistic that uses a different kernel function for each response variable, with a method for selecting each kernel function individually
• A method of adjusting for covariate effects on the response variables
• All significant computations performed using compiled C++ routines

Installation

You can install the development version of AMKAT using the following commands in R:

install.packages("devtools")
devtools::install_github("brianpatrickneal/AMKAT")

Example

Testing is straightforward using the amkat function. Below, we generate toy data for a set of feature variables (x) and for a set of response variables (y), as well as for a set of covariates (w), and test for joint association between the features and response:

library(AMKAT)
set.seed(42)
n <- 20 # sample size
y <- matrix(rnorm(3 * n), nrow = n, ncol = 3) # 3 response variables
x <- matrix(rnorm(50 * n), nrow = n, ncol = 50) # 50 feature variables
w <- matrix(rnorm(2 * n), nrow = n, ncol = 2) # 2 covariates
test_results <- amkat(y, x, covariates = w)
str(test_results)

Other Information

More details on the main function amkat can be found in its help file. Type ?AMKATpackage for an index of the other contents in the package's namespace.

brianpatrickneal/AMKAT documentation built on June 15, 2022, 8:47 a.m.