Iterative variable selection for high-dimensional data: prediction of pathological response in triple-negative breast cancer

In the last decade, regularized regression methods have offered alternatives forperforming multi-marker analysis and feature selection in a whole genome context.The process of defining a list of genes that will characterize an expressionprofile, remains unclear. This procedure oscillates between selecting the genes or transcripts of interest based on previous clinical evidence, or performing a whole transcriptome analys is that rests on advanced statistics. This paper introduces a methodology to deal with the variable selection and model estimation problems in the high-dimensional set-up, which can be particularly useful in the whole genome context. Results are validated using simulated data, and a real dataset from a triple negative breast cancer study.