Multiblock analysis of omics and imaging data with variable selection

Abstract : Sparse generalized canonical correlation analysis (SGCCA) has been proposed to combine RGCCA with an ℓ 1-penalty in a unified framework. Within this framework, blocks are not necessarily fully connected, which provides flexibility. The versatility and usefulness of SGCCA are illustrated on a 3-block dataset which combine Gene Expression, Comparative Genomic Hybridiza-tion and tumor location, determined on RMI at diagnosis. All data were measured on a cohort of 53 children with High Grade Glioma. SGCCA is available on CRAN as part of the RGCCA package.