Automated calibration for stability selection in penalised regression and graphical models

• DOI: 10.1093/jrsssc/qlad058
• Publication Date: 2023-07-13T16:37:44Z
• Authors: Barbara Bodinier; Sarah Filippi; Therese Haugdahl ...; Julien Chiquet; Marc Chadeau-Hyam
• Keywords: Lasso; Elastic net regularization; Graphical model

Abstract Stability selection represents an attractive approach to identify sparse sets of features jointly associated with outcome in high-dimensional contexts. We introduce automated calibration procedure via maximisation in-house stability score and accommodating a priori-known block structure (e.g. multi-OMIC) data. It applies [Least Absolute Shrinkage Selection Operator (LASSO)] penalised regression graphical models. Simulations show our outperforms non-stability-based approaches using the original calibration. Application multi-block LASSO on real (epigenetic transcriptomic) data from Norwegian Women Cancer study reveals central/credible novel cross-OMIC role LRRN3 biological response smoking. Proposed were implemented R package sharp.