Semi‐supervised empirical Bayes group‐regularized factor regression

Biometrical Journal, vol. 64, pp. 1289–1306

Abstract

The features in a high-dimensional biomedical prediction problem are often well described by low-dimensional latent variables (or factors). We use this to include unlabeled features and additional information on the features when building a prediction model. Such additional feature information is often available in biomedical applications. Examples are annotation of genes, metabolites, or p-values from a previous study. We employ a Bayesian factor regression model that jointly models the features and the outcome using Gaussian latent variables. We fit the model using a computationally efficient variational Bayes method, which scales to high dimensions. We use the extra information to set up a prior model for the features in terms of hyperparameters, which are then estimated through empirical Bayes. The method is demonstrated in simulations and two applications. One application considers influenza vaccine efficacy prediction based on microarray data. The second application predicts oral cancer metastasis from RNAseq data.

Authors 4

1. Magnus Münch Aachen Mathematical Institute

   Leiden University · Amsterdam University Medical Centers

   Affiliation as printed

   Department of Epidemiology & Data Science Amsterdam UMC Amsterdam The Netherlands

   Mathematical Institute Leiden University Leiden The Netherlands

   Mathematical Institute, Leiden University, Leiden, The Netherlands

2. Mark A. van de Wiel

   MRC Biostatistics Unit · Amsterdam University Medical Centers

   Affiliation as printed

   Department of Epidemiology & Data Science Amsterdam UMC Amsterdam The Netherlands

   MRC Biostatistics Unit Cambridge Institute of Public Health Cambridge UK

   MRC Biostatistics Unit, Cambridge Institute of Public Health, Cambridge, UK

3. Aad van der Vaart

   Delft University of Technology

   Affiliation as printed

   Delft Institute of Applied Mathematics ‐ TU Delft Delft The Netherlands

   Delft Institute of Applied Mathematics - TU Delft, Delft, The Netherlands

4. Carel F.W. Peeters corresponding

   Amsterdam University Medical Centers · Wageningen University & Research

   Affiliation as printed

   Department of Epidemiology & Data Science Amsterdam UMC Amsterdam The Netherlands

   Mathematical & Statistical Methods Group (Biometris) Wageningen University & Research Wageningen The Netherlands