Gene network reconstruction using global-local shrinkage priors

Gwenaël G.R. Leday, Mathisca C.M. De Gunst, Gino B. Kpogbezan, Aad W. Van Der Vaart, Wessel N. Van Wieringen, Mark A. Van De Wiel

Research output: Contribution to journalArticleAcademicpeer-review

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 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 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 ShrinkNet, is then applied to Glioblastoma to investigate the interactions between genes associated with patient survival.

Original languageEnglish
Pages (from-to)41-68
Number of pages28
JournalAnnals of Applied Statistics
Volume11
Issue number1
DOIs
Publication statusPublished - 1 Mar 2017

Cite this

Leday, Gwenaël G.R. ; De Gunst, Mathisca C.M. ; Kpogbezan, Gino B. ; Van Der Vaart, Aad W. ; Van Wieringen, Wessel N. ; Van De Wiel, Mark A. / Gene network reconstruction using global-local shrinkage priors. In: Annals of Applied Statistics. 2017 ; Vol. 11, No. 1. pp. 41-68.
@article{cf57e79c9c204c8b8fc0e97afa9309be,
title = "Gene network reconstruction using global-local shrinkage priors",
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 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 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 ShrinkNet, is then applied to Glioblastoma to investigate the interactions between genes associated with patient survival.",
keywords = "Bayesian inference, Empirical Bayes, Shrinkage, Undirected gene network, Variational approximation",
author = "Leday, {Gwena{\"e}l G.R.} and {De Gunst}, {Mathisca C.M.} and Kpogbezan, {Gino B.} and {Van Der Vaart}, {Aad W.} and {Van Wieringen}, {Wessel N.} and {Van De Wiel}, {Mark A.}",
year = "2017",
month = "3",
day = "1",
doi = "10.1214/16-AOAS990",
language = "English",
volume = "11",
pages = "41--68",
journal = "The Annals of Applied Statistics",
issn = "1932-6157",
publisher = "Institute of Mathematical Statistics",
number = "1",

}

Gene network reconstruction using global-local shrinkage priors. / Leday, Gwenaël G.R.; De Gunst, Mathisca C.M.; Kpogbezan, Gino B.; Van Der Vaart, Aad W.; Van Wieringen, Wessel N.; Van De Wiel, Mark A.

In: Annals of Applied Statistics, Vol. 11, No. 1, 01.03.2017, p. 41-68.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Gene network reconstruction using global-local shrinkage priors

AU - Leday, Gwenaël G.R.

AU - De Gunst, Mathisca C.M.

AU - Kpogbezan, Gino B.

AU - Van Der Vaart, Aad W.

AU - Van Wieringen, Wessel N.

AU - Van De Wiel, Mark A.

PY - 2017/3/1

Y1 - 2017/3/1

N2 - 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 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 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 ShrinkNet, is then applied to Glioblastoma to investigate the interactions between genes associated with patient survival.

AB - 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 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 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 ShrinkNet, is then applied to Glioblastoma to investigate the interactions between genes associated with patient survival.

KW - Bayesian inference

KW - Empirical Bayes

KW - Shrinkage

KW - Undirected gene network

KW - Variational approximation

UR - http://www.scopus.com/inward/record.url?scp=85017303881&partnerID=8YFLogxK

U2 - 10.1214/16-AOAS990

DO - 10.1214/16-AOAS990

M3 - Article

VL - 11

SP - 41

EP - 68

JO - The Annals of Applied Statistics

JF - The Annals of Applied Statistics

SN - 1932-6157

IS - 1

ER -