\name{mvtests} \alias{mvtests} \title{Multivariate SNP tests } \description{ This function calculates multivariate score tests between a multivariate (or multinomial) phenotype and sets of SNPs } \usage{ mvtests(phenotype, sets, stratum, data = sys.parent(), snp.data, rules = NULL, complete = TRUE, uncertain = FALSE, score = FALSE) } \arguments{ \item{phenotype}{ Either a factor (for a multinomial phenotype) or a matrix (for a multivariate phenotype) } \item{sets}{ A list of sets of SNPs to be tested against the phenotype } \item{stratum}{ (Optional) a stratifying variable } \item{data}{ A data frame in which \code{phenotype} and \code{stratum} reside. If absent these are assumed to be in the parent frame and correctly aligned with the rows of \code{snp.data} } \item{snp.data}{ An object of class \code{SnpMatrix} containing the SNP data } \item{rules}{ (Optional) A set of imputation rules. The function then carries out tess on imputed SNPs } \item{complete}{ If \code{TRUE} each test will use only subjects who have complete data for the phenotype and all SNPs in the set to be tested. If \code{FALSE}, then complete data for the phenotype is required, but tests are based upon complete pairs of SNPs } \item{uncertain}{ If \code{TRUE}, uncertain genotype calls will be used in the tests (scored by their posterior expectations). Otherwise such calls are treated as missing } \item{score}{ If \code{TRUE}, the score vectors and their variance-covariance matrices are saved in the output object for further processing } } \details{ Currently \code{complete=FALSE} is not implemented } \value{ An object of class \code{\link[=GlmTests-class]{snp.tests.glm}} or \code{\link[=GlmTestsScore-class]{GlmTests.score}} depending on whether \code{score} is set to \code{FALSE} or \code{TRUE} in the call } \author{David Clayton \email{david.clayton@cimr.cam.ac.uk}} \note{ This is an experimental version } \examples{ ## No example yet } \keyword{htest}