Implements several methods for testing the variance component parameter in regression models that contain kernel-based random effects, including a maximum of adjusted scores test. Several kernels are supported, including a profile hidden Markov model mutual information kernel for protein sequence. This package is described in Fong et al. (2015) .