This is an R package for conducting gene-based and pathway analysis. The current version supports both binary (logistic regression) and quantitative (linear regression) outcomes.
This package contains a set of three gene-based tests: AdaJoint, AdaJoint2, and ARTP. Details on these three tests are given in Zhang et al. (2013). It also implements a function for pathway analysis that uses the adaptive rank truncated product method (Yu et al, 2009) to combine gene-level p-values, which can be obtained by AdaJoint, AdaJoint2, or ARTP. Significant levels of gene and pathway analyses are evaluated through the computationally efficient direct simulation approach that generates statistics under the null hypothesis through asymptotic multivariate normal distributions. This package has two major improvements over the ARTP package: 1. it provides more options for conducting gene and pathway analyses; and 2. it runs much faster as it uses a computationally efficient algorithm (based on asymptotic distributions) to evaluate pvalues, instead of a permutation procedure, which is used by the ARTP package.
Kai Yu and Han Zhang