A scale-free structure prior for Bayesian inference of Gaussian graphical models

Osamu Maruyama, Shota Shikita

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

The inference of gene association networks from gene expression profiles is an important approach to elucidate various cellular mechanisms. However, there exists a problematic issue that the number of samples is relatively small than that of genes. A promising approach to this problem will be to design regularization terms for characteristic network structures like sparsity and scale-freeness and optimize a scoring function including those regularization terms. The inference problem for gene association networks is often formulated as the problem of estimating the inverse covariance matrix of a Gaussian distribution from its samples. For this Bayesian inference problem, we propose a novel scale-free structure prior and devise a sampling method for optimizing a posterior probability including the prior. In a simulation study, scale-free graphs of 30 and 100 nodes are generated by the Barabási-Albert model, and the proposed method is shown to outperform another method which also use a scale-free regularization term. Our method is also applied to real gene expression profiles, and the resulting graph shows biologically meaningful features. Thus, we empirically conclude that our scale-free structure prior is effective in Bayesian inference of Gaussian graphical models.

Original languageEnglish
Title of host publicationProceedings - 2014 IEEE International Conference on Bioinformatics and Biomedicine, IEEE BIBM 2014
EditorsHuiru Zheng, Xiaohua Tony Hu, Yadong Wang, Jin-Kao Hao, David Gilbert, Daniel Berrar, Kwang-Hyun Cho, Werner Dubitzky
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages131-138
Number of pages8
ISBN (Electronic)9781479956692
DOIs
Publication statusPublished - Dec 29 2014
Event2014 IEEE International Conference on Bioinformatics and Biomedicine, IEEE BIBM 2014 - Belfast, United Kingdom
Duration: Nov 2 2014Nov 5 2014

Publication series

NameProceedings - 2014 IEEE International Conference on Bioinformatics and Biomedicine, IEEE BIBM 2014

Other

Other2014 IEEE International Conference on Bioinformatics and Biomedicine, IEEE BIBM 2014
CountryUnited Kingdom
CityBelfast
Period11/2/1411/5/14

Fingerprint

Genes
Gene expression
Gene Regulatory Networks
Transcriptome
Gaussian distribution
Covariance matrix
Normal Distribution
Sampling

All Science Journal Classification (ASJC) codes

  • Biomedical Engineering
  • Health Informatics

Cite this

Maruyama, O., & Shikita, S. (2014). A scale-free structure prior for Bayesian inference of Gaussian graphical models. In H. Zheng, X. T. Hu, Y. Wang, J-K. Hao, D. Gilbert, D. Berrar, K-H. Cho, ... W. Dubitzky (Eds.), Proceedings - 2014 IEEE International Conference on Bioinformatics and Biomedicine, IEEE BIBM 2014 (pp. 131-138). [6999141] (Proceedings - 2014 IEEE International Conference on Bioinformatics and Biomedicine, IEEE BIBM 2014). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/BIBM.2014.6999141

A scale-free structure prior for Bayesian inference of Gaussian graphical models. / Maruyama, Osamu; Shikita, Shota.

Proceedings - 2014 IEEE International Conference on Bioinformatics and Biomedicine, IEEE BIBM 2014. ed. / Huiru Zheng; Xiaohua Tony Hu; Yadong Wang; Jin-Kao Hao; David Gilbert; Daniel Berrar; Kwang-Hyun Cho; Werner Dubitzky. Institute of Electrical and Electronics Engineers Inc., 2014. p. 131-138 6999141 (Proceedings - 2014 IEEE International Conference on Bioinformatics and Biomedicine, IEEE BIBM 2014).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Maruyama, O & Shikita, S 2014, A scale-free structure prior for Bayesian inference of Gaussian graphical models. in H Zheng, XT Hu, Y Wang, J-K Hao, D Gilbert, D Berrar, K-H Cho & W Dubitzky (eds), Proceedings - 2014 IEEE International Conference on Bioinformatics and Biomedicine, IEEE BIBM 2014., 6999141, Proceedings - 2014 IEEE International Conference on Bioinformatics and Biomedicine, IEEE BIBM 2014, Institute of Electrical and Electronics Engineers Inc., pp. 131-138, 2014 IEEE International Conference on Bioinformatics and Biomedicine, IEEE BIBM 2014, Belfast, United Kingdom, 11/2/14. https://doi.org/10.1109/BIBM.2014.6999141
Maruyama O, Shikita S. A scale-free structure prior for Bayesian inference of Gaussian graphical models. In Zheng H, Hu XT, Wang Y, Hao J-K, Gilbert D, Berrar D, Cho K-H, Dubitzky W, editors, Proceedings - 2014 IEEE International Conference on Bioinformatics and Biomedicine, IEEE BIBM 2014. Institute of Electrical and Electronics Engineers Inc. 2014. p. 131-138. 6999141. (Proceedings - 2014 IEEE International Conference on Bioinformatics and Biomedicine, IEEE BIBM 2014). https://doi.org/10.1109/BIBM.2014.6999141
Maruyama, Osamu ; Shikita, Shota. / A scale-free structure prior for Bayesian inference of Gaussian graphical models. Proceedings - 2014 IEEE International Conference on Bioinformatics and Biomedicine, IEEE BIBM 2014. editor / Huiru Zheng ; Xiaohua Tony Hu ; Yadong Wang ; Jin-Kao Hao ; David Gilbert ; Daniel Berrar ; Kwang-Hyun Cho ; Werner Dubitzky. Institute of Electrical and Electronics Engineers Inc., 2014. pp. 131-138 (Proceedings - 2014 IEEE International Conference on Bioinformatics and Biomedicine, IEEE BIBM 2014).
@inproceedings{0365c8ff381c447cb1661a6d48df2d76,
title = "A scale-free structure prior for Bayesian inference of Gaussian graphical models",
abstract = "The inference of gene association networks from gene expression profiles is an important approach to elucidate various cellular mechanisms. However, there exists a problematic issue that the number of samples is relatively small than that of genes. A promising approach to this problem will be to design regularization terms for characteristic network structures like sparsity and scale-freeness and optimize a scoring function including those regularization terms. The inference problem for gene association networks is often formulated as the problem of estimating the inverse covariance matrix of a Gaussian distribution from its samples. For this Bayesian inference problem, we propose a novel scale-free structure prior and devise a sampling method for optimizing a posterior probability including the prior. In a simulation study, scale-free graphs of 30 and 100 nodes are generated by the Barab{\'a}si-Albert model, and the proposed method is shown to outperform another method which also use a scale-free regularization term. Our method is also applied to real gene expression profiles, and the resulting graph shows biologically meaningful features. Thus, we empirically conclude that our scale-free structure prior is effective in Bayesian inference of Gaussian graphical models.",
author = "Osamu Maruyama and Shota Shikita",
year = "2014",
month = "12",
day = "29",
doi = "10.1109/BIBM.2014.6999141",
language = "English",
series = "Proceedings - 2014 IEEE International Conference on Bioinformatics and Biomedicine, IEEE BIBM 2014",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "131--138",
editor = "Huiru Zheng and Hu, {Xiaohua Tony} and Yadong Wang and Jin-Kao Hao and David Gilbert and Daniel Berrar and Kwang-Hyun Cho and Werner Dubitzky",
booktitle = "Proceedings - 2014 IEEE International Conference on Bioinformatics and Biomedicine, IEEE BIBM 2014",
address = "United States",

}

TY - GEN

T1 - A scale-free structure prior for Bayesian inference of Gaussian graphical models

AU - Maruyama, Osamu

AU - Shikita, Shota

PY - 2014/12/29

Y1 - 2014/12/29

N2 - The inference of gene association networks from gene expression profiles is an important approach to elucidate various cellular mechanisms. However, there exists a problematic issue that the number of samples is relatively small than that of genes. A promising approach to this problem will be to design regularization terms for characteristic network structures like sparsity and scale-freeness and optimize a scoring function including those regularization terms. The inference problem for gene association networks is often formulated as the problem of estimating the inverse covariance matrix of a Gaussian distribution from its samples. For this Bayesian inference problem, we propose a novel scale-free structure prior and devise a sampling method for optimizing a posterior probability including the prior. In a simulation study, scale-free graphs of 30 and 100 nodes are generated by the Barabási-Albert model, and the proposed method is shown to outperform another method which also use a scale-free regularization term. Our method is also applied to real gene expression profiles, and the resulting graph shows biologically meaningful features. Thus, we empirically conclude that our scale-free structure prior is effective in Bayesian inference of Gaussian graphical models.

AB - The inference of gene association networks from gene expression profiles is an important approach to elucidate various cellular mechanisms. However, there exists a problematic issue that the number of samples is relatively small than that of genes. A promising approach to this problem will be to design regularization terms for characteristic network structures like sparsity and scale-freeness and optimize a scoring function including those regularization terms. The inference problem for gene association networks is often formulated as the problem of estimating the inverse covariance matrix of a Gaussian distribution from its samples. For this Bayesian inference problem, we propose a novel scale-free structure prior and devise a sampling method for optimizing a posterior probability including the prior. In a simulation study, scale-free graphs of 30 and 100 nodes are generated by the Barabási-Albert model, and the proposed method is shown to outperform another method which also use a scale-free regularization term. Our method is also applied to real gene expression profiles, and the resulting graph shows biologically meaningful features. Thus, we empirically conclude that our scale-free structure prior is effective in Bayesian inference of Gaussian graphical models.

UR - http://www.scopus.com/inward/record.url?scp=84922794338&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84922794338&partnerID=8YFLogxK

U2 - 10.1109/BIBM.2014.6999141

DO - 10.1109/BIBM.2014.6999141

M3 - Conference contribution

AN - SCOPUS:84922794338

T3 - Proceedings - 2014 IEEE International Conference on Bioinformatics and Biomedicine, IEEE BIBM 2014

SP - 131

EP - 138

BT - Proceedings - 2014 IEEE International Conference on Bioinformatics and Biomedicine, IEEE BIBM 2014

A2 - Zheng, Huiru

A2 - Hu, Xiaohua Tony

A2 - Wang, Yadong

A2 - Hao, Jin-Kao

A2 - Gilbert, David

A2 - Berrar, Daniel

A2 - Cho, Kwang-Hyun

A2 - Dubitzky, Werner

PB - Institute of Electrical and Electronics Engineers Inc.

ER -