An empirical Bayes approach to network recovery using external knowledge

Gino B. Kpogbezan, Aad W. van der Vaart, Wessel N. van Wieringen, Gwenaël G.R. Leday, Mark A. van de Wiel

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

Reconstruction of a high-dimensional network may benefit substantially from the inclusion of prior knowledge on the network topology. In the case of gene interaction networks such knowledge may come for instance from pathway repositories like KEGG, or be inferred from data of a pilot study. The Bayesian framework provides a natural means of including such prior knowledge. Based on a Bayesian Simultaneous Equation Model, we develop an appealing Empirical Bayes (EB) procedure that automatically assesses the agreement of the used prior knowledge with the data at hand. We use variational Bayes method for posterior densities approximation and compare its accuracy with that of Gibbs sampling strategy. Our method is computationally fast, and can outperform known competitors. In a simulation study, we show that accurate prior data can greatly improve the reconstruction of the network, but need not harm the reconstruction if wrong. We demonstrate the benefits of the method in an analysis of gene expression data from GEO. In particular, the edges of the recovered network have superior reproducibility (compared to that of competitors) over resampled versions of the data.

Original languageEnglish
Pages (from-to)932-947
Number of pages16
JournalBiometrical Journal
Volume59
Issue number5
DOIs
Publication statusPublished - 1 Sep 2017

Cite this

Kpogbezan, Gino B. ; van der Vaart, Aad W. ; van Wieringen, Wessel N. ; Leday, Gwenaël G.R. ; van de Wiel, Mark A. / An empirical Bayes approach to network recovery using external knowledge. In: Biometrical Journal. 2017 ; Vol. 59, No. 5. pp. 932-947.
@article{2de70d5cd6a64586a7d47e8d30550fbc,
title = "An empirical Bayes approach to network recovery using external knowledge",
abstract = "Reconstruction of a high-dimensional network may benefit substantially from the inclusion of prior knowledge on the network topology. In the case of gene interaction networks such knowledge may come for instance from pathway repositories like KEGG, or be inferred from data of a pilot study. The Bayesian framework provides a natural means of including such prior knowledge. Based on a Bayesian Simultaneous Equation Model, we develop an appealing Empirical Bayes (EB) procedure that automatically assesses the agreement of the used prior knowledge with the data at hand. We use variational Bayes method for posterior densities approximation and compare its accuracy with that of Gibbs sampling strategy. Our method is computationally fast, and can outperform known competitors. In a simulation study, we show that accurate prior data can greatly improve the reconstruction of the network, but need not harm the reconstruction if wrong. We demonstrate the benefits of the method in an analysis of gene expression data from GEO. In particular, the edges of the recovered network have superior reproducibility (compared to that of competitors) over resampled versions of the data.",
keywords = "Empirical Bayes, High-dimensional Bayesian inference, Prior information, Undirected network, Variational approximation",
author = "Kpogbezan, {Gino B.} and {van der Vaart}, {Aad W.} and {van Wieringen}, {Wessel N.} and Leday, {Gwena{\"e}l G.R.} and {van de Wiel}, {Mark A.}",
year = "2017",
month = "9",
day = "1",
doi = "10.1002/bimj.201600090",
language = "English",
volume = "59",
pages = "932--947",
journal = "Biometrical Journal",
issn = "0323-3847",
publisher = "Wiley-VCH Verlag",
number = "5",

}

An empirical Bayes approach to network recovery using external knowledge. / Kpogbezan, Gino B.; van der Vaart, Aad W.; van Wieringen, Wessel N.; Leday, Gwenaël G.R.; van de Wiel, Mark A.

In: Biometrical Journal, Vol. 59, No. 5, 01.09.2017, p. 932-947.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - An empirical Bayes approach to network recovery using external knowledge

AU - Kpogbezan, Gino B.

AU - van der Vaart, Aad W.

AU - van Wieringen, Wessel N.

AU - Leday, Gwenaël G.R.

AU - van de Wiel, Mark A.

PY - 2017/9/1

Y1 - 2017/9/1

N2 - Reconstruction of a high-dimensional network may benefit substantially from the inclusion of prior knowledge on the network topology. In the case of gene interaction networks such knowledge may come for instance from pathway repositories like KEGG, or be inferred from data of a pilot study. The Bayesian framework provides a natural means of including such prior knowledge. Based on a Bayesian Simultaneous Equation Model, we develop an appealing Empirical Bayes (EB) procedure that automatically assesses the agreement of the used prior knowledge with the data at hand. We use variational Bayes method for posterior densities approximation and compare its accuracy with that of Gibbs sampling strategy. Our method is computationally fast, and can outperform known competitors. In a simulation study, we show that accurate prior data can greatly improve the reconstruction of the network, but need not harm the reconstruction if wrong. We demonstrate the benefits of the method in an analysis of gene expression data from GEO. In particular, the edges of the recovered network have superior reproducibility (compared to that of competitors) over resampled versions of the data.

AB - Reconstruction of a high-dimensional network may benefit substantially from the inclusion of prior knowledge on the network topology. In the case of gene interaction networks such knowledge may come for instance from pathway repositories like KEGG, or be inferred from data of a pilot study. The Bayesian framework provides a natural means of including such prior knowledge. Based on a Bayesian Simultaneous Equation Model, we develop an appealing Empirical Bayes (EB) procedure that automatically assesses the agreement of the used prior knowledge with the data at hand. We use variational Bayes method for posterior densities approximation and compare its accuracy with that of Gibbs sampling strategy. Our method is computationally fast, and can outperform known competitors. In a simulation study, we show that accurate prior data can greatly improve the reconstruction of the network, but need not harm the reconstruction if wrong. We demonstrate the benefits of the method in an analysis of gene expression data from GEO. In particular, the edges of the recovered network have superior reproducibility (compared to that of competitors) over resampled versions of the data.

KW - Empirical Bayes

KW - High-dimensional Bayesian inference

KW - Prior information

KW - Undirected network

KW - Variational approximation

UR - http://www.scopus.com/inward/record.url?scp=85017477984&partnerID=8YFLogxK

U2 - 10.1002/bimj.201600090

DO - 10.1002/bimj.201600090

M3 - Article

VL - 59

SP - 932

EP - 947

JO - Biometrical Journal

JF - Biometrical Journal

SN - 0323-3847

IS - 5

ER -