RVTESTS takes standard VCF/BCF (for genotypes) and PED (for phenotypes and covariates) files as input. We implement frequently used rare variant association tests, including burden tests, SKAT and variable threshold tests (and many others). These tests allow for the analyses of both autosomal and X-linked genes. We also extend these methods to the analysis of (cryptically) related individuals using Linear Mixed Models (LMMs) (with both empirical and pedigree kinship). To enable the analysis of large datasets with many thousands of individuals genotyped on millions of markers, we incorporate recently developed efficient algorithms for fitting LMMs (see Supplementary Material). Given that rare variant association analyses may be distinctly confounded by population structures, linear mixed model analysis has been identified as a key approach for the proper control of type I errors in rare variant association analyses (Lippert et al., 2011; Listgarten et al., 2013). LMM analyses supported by RVTESTS can be quite valuable.
