We propose a multivariate sparse group lasso variable selection and estimation method for data with highdimensional predictors as well as high-dimensional response variables. The method is carried out ...