Gene Network Reconstruction using Global-Local Shrinkage Priors
Authors
De Gunst, MCM
Krogbezan, GB
van der Vaart, AW
van Wieringen, WN
van de Wiel, MA
Publication Date
2017-04-08Journal Title
The Annals of Applied Statistics
ISSN
1932-6157
Publisher
Institute of Mathematical Statistics
Volume
11
Issue
1
Pages
41-68
Language
English
Type
Article
This Version
VoR
Metadata
Show full item recordCitation
Leday, G., De Gunst, M., Krogbezan, G., van der Vaart, A., van Wieringen, W., & van de Wiel, M. (2017). Gene Network Reconstruction using Global-Local Shrinkage Priors. The Annals of Applied Statistics, 11 (1), 41-68. https://doi.org/10.1214/16-AOAS990
Abstract
Reconstructing a gene network from high-throughput molecular data is an important but challenging task, as the number of parameters to estimate easily is much larger than the sample size. A conventional remedy is to regularize or penalize the model likelihood. In network models, this is often done $\textit{locally}$ in the neighborhood of each node or gene. However, estimation of the many regularization parameters is often difficult and can result in large statistical uncertainties. In this paper we propose to combine local regularization with $\textit{global}$ shrinkage of the regularization parameters to borrow strength between genes and improve inference. We employ a simple Bayesian model with nonsparse, conjugate priors to facilitate the use of fast variational approximations to posteriors. We discuss empirical Bayes estimation of hyperparameters of the priors, and propose a novel approach to rank-based posterior thresholding. Using extensive model- and data-based simulations, we demonstrate that the proposed inference strategy outperforms popular (sparse) methods, yields more stable edges, and is more reproducible. The proposed method, termed $\texttt{ShrinkNet}$, is then applied to Glioblastoma to investigate the interactions between genes associated with patient survival.
Keywords
undirected gene network, Bayesian inference, shrinkage, variational approximation, empirical Bayes
Sponsorship
This work was supported by the Center for Medical Systems Biology (CMSB), and the European Union Grant EpiRadBio, established by the Netherlands Genomics Initiative/Netherlands Organization for Scientific Research (NGI/NWO), nr. FP7-269553.
Embargo Lift Date
2100-01-01
Identifiers
External DOI: https://doi.org/10.1214/16-AOAS990
This record's URL: https://www.repository.cam.ac.uk/handle/1810/262410
Rights
Licence:
http://www.rioxx.net/licenses/all-rights-reserved