## Combining microarrays and biological knowledge for estimating gene networks via Bayesian networks (2003)

### Cached

### Download Links

- [bonsai.ims.u-tokyo.ac.jp]
- [bonsai.hgc.jp]
- [conferences.computer.org]
- [bonsai.ims.u-tokyo.ac.jp]
- [bonsai.hgc.jp]
- [tswww.ism.ac.jp]
- DBLP

### Other Repositories/Bibliography

Venue: | In Proceedings of the IEEE Computer Society Bioinformatics Conference (CSB 03 |

Citations: | 59 - 5 self |

### BibTeX

@INPROCEEDINGS{Imoto03combiningmicroarrays,

author = {Seiya Imoto and Tomoyuki Higuchi and Takao Goto and Kousuke Tashiro and Satoru Kuhara and Satoru Miyano},

title = {Combining microarrays and biological knowledge for estimating gene networks via Bayesian networks},

booktitle = {In Proceedings of the IEEE Computer Society Bioinformatics Conference (CSB 03},

year = {2003},

pages = {104--113},

publisher = {IEEE}

}

### OpenURL

### Abstract

We propose a statistical method for estimating a gene network based on Bayesian networks from microarray gene expression data together with biological knowledge including protein-protein interactions, protein-DNA interactions, binding site information, existing literature and so on. Unfortunately, microarray data do not contain enough information for constructing gene networks accurately in many cases. Our method adds biological knowledge to the estimation method of gene networks under a Bayesian statistical framework, and also controls the trade-off between microarray information and biological knowledge automatically. We conduct Monte Carlo simulations to show the effectiveness of the proposed method. We analyze Saccharomyces cerevisiae gene expression data as an application. 1.

### Citations

4020 |
Stochastic relaxation, Gibbs distributions, and the Bayesian restoration of images
- Geman, Geman
- 1984
(Show Context)
Citation Context ...�� is a local energy defined by gene� and its parents. Figure 1 shows an example of a gene network and its energy. The probability of a network �, � � , is naturally modeled by the Gibbs distribution =-=[15]-=- � � ��s�ÜÔ�s�� � �� (2) where � � is a hyperparameter and � is a normalizing constant called the partition function � � � � � �ÜÔ�s�� � �� Here � is the set of possible networks. By replacing �À � ��... |

1613 | Generalized Additive Models
- Hastie, Tibshirani
- 1990
(Show Context)
Citation Context ...hereÊ�� and ��� are normalized intensities of Cy5 and Cy3 for gene� measured by �th microarray. The interaction between gene� and its parents is modeled by the nonparametric additive regression model =-=[19]-=- with heterogeneous error variances Ü�� � Ñ� Ô � � ¡¡¡ Ñ�Õ� Ô � �Õ� ���� where Ô � �� is the expression value of �th parent of gene� measured by �th microarray and ��� depends independently Proceeding... |

1259 |
A Practical Guide to Splines
- Boor
- 1978
(Show Context)
Citation Context ...roceedings of the Computational Systems Bioinformatics (CSB’03) 0-7695-2000-6/03 $17.00 © 2003 IEEE and normally on mean 0 and variance � ��. Here, Ñ�� ¡ is a smooth function constructed by �-splines =-=[9, 12, 24]-=- of the form where �� � � Ñ�� Ô � �� � Å�� � Ñ� � � � Ñ��Ñ� Ô�� � � ¡ ������Å���� ¡ � is a prescribed set of �- splines and � Ñ� are parameters. Hence, a Bayesian network and nonparametric heteros... |

1132 | A Bayesian method for the induction of probabilistic networks from data - Cooper, Herskovits - 1992 |

969 |
An introduction to Bayesian Networks
- Jensen
- 1996
(Show Context)
Citation Context ...thod to Saccharomyces cerevisiae gene expression data in Section 3.2. 2. Method for Estimating Gene Networks 2.1. Bayesian network and nonparametric heteroscedastic regression model Bayesian networks =-=[26]-=- are a type of graphical models for capturing complex relationships among a large amount of random variables by the directed acyclic graph encoding the Markov assumption. In the context of Bayesian ne... |

895 | A tutorial on learning with Bayesian Networks
- Heckerman
- 1995
(Show Context)
Citation Context ...oolean networks [1, 2, 32, 42], differenProceedings of the Computational Systems Bioinformatics (CSB’03) 0-7695-2000-6/03 $17.00 © 2003 IEEE tial equation models [7, 10, 11, 32] and Bayesian networks =-=[13, 14, 17, 18, 20, 22, 23, 37]-=-. Main drawback for the gene network construction from microarray data is that while the gene network contains a large number of genes, the information contained in gene expression data is limited by ... |

800 | Using bayesian networks to analyze expression data
- Friedman
- 2000
(Show Context)
Citation Context ...oolean networks [1, 2, 32, 42], differenProceedings of the Computational Systems Bioinformatics (CSB’03) 0-7695-2000-6/03 $17.00 © 2003 IEEE tial equation models [7, 10, 11, 32] and Bayesian networks =-=[13, 14, 17, 18, 20, 22, 23, 37]-=-. Main drawback for the gene network construction from microarray data is that while the gene network contains a large number of genes, the information contained in gene expression data is limited by ... |

713 |
A comprehensive two-hybrid analysis to explore the yeast protein interactome
- Ito, Chiba, et al.
- 2001
(Show Context)
Citation Context ... In particular, the direction of gene regulation is difficult to decide using gene expression data only. Hence, the use of biological knowledge, including protein-protein and protein-DNA interactions =-=[3, 5, 16, 21, 25]-=-, sequences of the binding site of the genes controlled by transcription regulators [31, 40, 47], literature and so on, are considered to be a key for microarray data analysis. The use of biological k... |

508 |
Transcriptional regulatory networks in saccharomyces cerevisiae
- Lee
- 2002
(Show Context)
Citation Context ...y. Hence, the use of biological knowledge, including protein-protein and protein-DNA interactions [3, 5, 16, 21, 25], sequences of the binding site of the genes controlled by transcription regulators =-=[31, 40, 47]-=-, literature and so on, are considered to be a key for microarray data analysis. The use of biological knowledge has previously received considerable attention for extracting more information from mic... |

292 |
Microarray data normalization and transformation
- Quackenbush
- 2002
(Show Context)
Citation Context ...ensional gene expression vector obtained by �th microarray. Here, Ü�� is an expression value of �th gene, denoted by gene�, measured by �th microarray after required normalizations and transformation =-=[39]-=-. Ordinary, Ü�� is given by ÐÓ� Ê������ ,whereÊ�� and ��� are normalized intensities of Cy5 and Cy3 for gene� measured by �th microarray. The interaction between gene� and its parents is modeled by th... |

290 |
Regulatory element detection using correlation with expression
- Bussemaker, Li, et al.
- 2001
(Show Context)
Citation Context ...and so on, are considered to be a key for microarray data analysis. The use of biological knowledge has previously received considerable attention for extracting more information from microarray data =-=[4, 6, 18, 33, 36, 38, 41]-=-. In this paper, we provide a general framework for combining microarray data and biological knowledge aimed at estimating a gene network by using a Bayesian network model. If the gene regulation mech... |

289 | BIND–the biomolecular interaction network database
- Bader, Donaldson, et al.
- 2001
(Show Context)
Citation Context ... In particular, the direction of gene regulation is difficult to decide using gene expression data only. Hence, the use of biological knowledge, including protein-protein and protein-DNA interactions =-=[3, 5, 16, 21, 25]-=-, sequences of the binding site of the genes controlled by transcription regulators [31, 40, 47], literature and so on, are considered to be a key for microarray data analysis. The use of biological k... |

254 | Probabilistic boolean networks: A rule-based uncertainty model for gene regulatory networks
- Shmulevich, Dougherty, et al.
- 2002
(Show Context)
Citation Context ...k has become one of the central topics in the field of bioinformatics. Several methodologies have been proposed for constructing a gene network based on gene expression data, such as Boolean networks =-=[1, 2, 32, 42]-=-, differenProceedings of the Computational Systems Bioinformatics (CSB’03) 0-7695-2000-6/03 $17.00 © 2003 IEEE tial equation models [7, 10, 11, 32] and Bayesian networks [13, 14, 17, 18, 20, 22, 23, 3... |

244 | Learning Bayesian networks with local structure
- Friedman, Goldszmidt
- 1996
(Show Context)
Citation Context ...oolean networks [1, 2, 32, 42], differenProceedings of the Computational Systems Bioinformatics (CSB’03) 0-7695-2000-6/03 $17.00 © 2003 IEEE tial equation models [7, 10, 11, 32] and Bayesian networks =-=[13, 14, 17, 18, 20, 22, 23, 37]-=-. Main drawback for the gene network construction from microarray data is that while the gene network contains a large number of genes, the information contained in gene expression data is limited by ... |

225 |
Accurate approximations for posterior moments and marginal densities JASA
- TIERNEY, KADANE
- 1986
(Show Context)
Citation Context ...terion based on the posterior probability of the network is how to compute the marginal likelihood given by a high dimensional integral. Imoto et al. [23] used the Laplace approximation for integrals =-=[8, 30, 45]-=- and derived a criterion, named BNRC��Ø�ÖÓ (Bayesian network and Nonparametric heteroscedastic Regression Criterion), of the form where �ÆÊ���Ø�ÖÓ � �sÐÓ�� � ÐÓ� ¬ ¬ Ò Ð� ���� � Ò �� Â� �� �s� Â� � ��... |

225 | Flexible smoothing with B-splines and penalties
- EILERS, MARX
- 1996
(Show Context)
Citation Context ...ssion value of kth parent of genej measured by ith microarray and εij depends independently iqj and normally on mean 0 and variance σ 2 ij . Here, mjk(·) is a smooth function constructed by B-splines =-=[9, 12, 24]-=- of the form mjk(p (j) Mjk ∑ ik ) = γ m=1 (j) mkb(j) mk (p(j) ik ), where {b (j) 1k (·), ..., b(j) Mjk,k (·)} is a prescribed set of Bsplines and γ (j) mk are parameters. Hence, a Bayesian network and... |

203 | Identification of genetic networks from a small number of gene expression patterns under the boolean network model
- Akutsu, Miyano, et al.
- 1999
(Show Context)
Citation Context ...k has become one of the central topics in the field of bioinformatics. Several methodologies have been proposed for constructing a gene network based on gene expression data, such as Boolean networks =-=[1, 2, 32, 42]-=-, differenProceedings of the Computational Systems Bioinformatics (CSB’03) 0-7695-2000-6/03 $17.00 © 2003 IEEE tial equation models [7, 10, 11, 32] and Bayesian networks [13, 14, 17, 18, 20, 22, 23, 3... |

197 |
Identifying regulatory networks by combinatorial analysis of promoter elements
- Pilpel
- 2001
(Show Context)
Citation Context ...and so on, are considered to be a key for microarray data analysis. The use of biological knowledge has previously received considerable attention for extracting more information from microarray data =-=[4, 6, 18, 33, 36, 38, 41]-=-. In this paper, we provide a general framework for combining microarray data and biological knowledge aimed at estimating a gene network by using a Bayesian network model. If the gene regulation mech... |

187 | Modeling gene expression with differential equations
- Chen, He, et al.
- 1999
(Show Context)
Citation Context ...sed on gene expression data, such as Boolean networks [1, 2, 32, 42], differenProceedings of the Computational Systems Bioinformatics (CSB’03) 0-7695-2000-6/03 $17.00 © 2003 IEEE tial equation models =-=[7, 10, 11, 32]-=- and Bayesian networks [13, 14, 17, 18, 20, 22, 23, 37]. Main drawback for the gene network construction from microarray data is that while the gene network contains a large number of genes, the infor... |

168 | Modeling gene expression data using dynamic bayesian networks
- Murphy, Mian
(Show Context)
Citation Context ...n in Imoto et al. [23]. Note that the proposed prior probabilityof the network can be used for other types of Bayesian network models, such as discrete Bayesian networks and dynamic Bayesian networks =-=[29, 34, 36, 43]-=-. The computation of partition function, �, is intractable even for moderately sized gene networks. To avoid this problem, we compute upper and lower bounds of the partial function and use them for ch... |

166 | Inferring subnetworks from perturbed expression profiles
- Pe’er, Regev, et al.
(Show Context)
Citation Context |

156 | Random Fields and Their Applications - Kindermann, Snell - 1980 |

139 | Using graphical models and genomic expression data to statistically validate models of genetic regulatory networks
- Hartemink, Gifford, et al.
- 2001
(Show Context)
Citation Context |

138 |
Discovering regulatory and signalling circuits in molecular interaction networks
- Ideker
- 2002
(Show Context)
Citation Context ... In particular, the direction of gene regulation is difficult to decide using gene expression data only. Hence, the use of biological knowledge, including protein-protein and protein-DNA interactions =-=[3, 5, 16, 21, 25]-=-, sequences of the binding site of the genes controlled by transcription regulators [31, 40, 47], literature and so on, are considered to be a key for microarray data analysis. The use of biological k... |

118 | Discovering Molecular Pathways from Protein Interaction and Gene Expression Data - Segal, Wang, et al. - 2003 |

103 | Combining location and expression data for principled discovery of genetic regulatory networks
- Hartemink, Gifford, et al.
- 2002
(Show Context)
Citation Context |

102 | Genome-wide discovery of transcriptional modules from DNA sequence and gene expression - Segal - 2003 |

101 |
Inferring qualitative relations in genetic networks and metabolic pathways
- Akutsu, Miyano, et al.
- 2000
(Show Context)
Citation Context ...k has become one of the central topics in the field of bioinformatics. Several methodologies have been proposed for constructing a gene network based on gene expression data, such as Boolean networks =-=[1, 2, 32, 42]-=-, differenProceedings of the Computational Systems Bioinformatics (CSB’03) 0-7695-2000-6/03 $17.00 © 2003 IEEE tial equation models [7, 10, 11, 32] and Bayesian networks [13, 14, 17, 18, 20, 22, 23, 3... |

98 | Miyano: Estimation of genetic networks and functional structures between genes by using Bayesian networks and nonparametric regression. Pac Symp Biocomput175-86
- Imoto, Goto, et al.
- 2002
(Show Context)
Citation Context |

65 | Inferring gene regulatory networks from time ordered gene expression data of Bacillus subtilis using differential equations
- Hoon, Imoto, et al.
- 2003
(Show Context)
Citation Context ...sed on gene expression data, such as Boolean networks [1, 2, 32, 42], differenProceedings of the Computational Systems Bioinformatics (CSB’03) 0-7695-2000-6/03 $17.00 © 2003 IEEE tial equation models =-=[7, 10, 11, 32]-=- and Bayesian networks [13, 14, 17, 18, 20, 22, 23, 37]. Main drawback for the gene network construction from microarray data is that while the gene network contains a large number of genes, the infor... |

63 |
Estimating gene networks from gene expression data by combining Bayesian network model with promoter element detection
- Tamada
- 2003
(Show Context)
Citation Context ...be introduced directly into our method. One straightforward way is the use of known regulatory motifs kept in public databases such as SCPD [40] and YTF [47]. As for an advanced method, Tamada et al. =-=[44]-=- proposed a method for simultaneously estimating a gene network and detecting regulatory motifs based on our method, and succeeded in estimating an accurate gene network and detecting a true regulator... |

62 | Development of a system for the inference of large scale genetic networks
- Maki, Tominaga, et al.
- 2001
(Show Context)
Citation Context |

59 | From Promoter Sequence to Expression: A Probabilistic Framework
- Segal, Barash, et al.
- 2002
(Show Context)
Citation Context ...and so on, are considered to be a key for microarray data analysis. The use of biological knowledge has previously received considerable attention for extracting more information from microarray data =-=[4, 6, 18, 33, 36, 38, 41]-=-. In this paper, we provide a general framework for combining microarray data and biological knowledge aimed at estimating a gene network by using a Bayesian network model. If the gene regulation mech... |

55 | Dynamic Bayesian network and nonparametric regression for nonlinear modeling of gene networks from time series gene expression data
- Kim, Imoto, et al.
- 2003
(Show Context)
Citation Context ...n in Imoto et al. [23]. Note that the proposed prior probabilityof the network can be used for other types of Bayesian network models, such as discrete Bayesian networks and dynamic Bayesian networks =-=[29, 34, 36, 43]-=-. The computation of partition function, �, is intractable even for moderately sized gene networks. To avoid this problem, we compute upper and lower bounds of the partial function and use them for ch... |

53 | Modelling regulatory pathways in E. coli from time series expression profiles
- Ong, Glasner, et al.
- 2002
(Show Context)
Citation Context |

48 | Inferring gene networks from time series microarray data using dynamic Bayesian networks - Kim, Imoto, et al. - 2003 |

38 | Bayesian network and nonparametric heteroscedastic regression for nonlinear modeling of genetic network
- Imoto, Kim, et al.
(Show Context)
Citation Context |

38 | Evaluating Functional Network Inference Using Simulations of Complex Biological Systems. (2002) Accepted by The 10th international conference on Intelligent Systems for Molecular Biology
- Smith, Jarvis, et al.
(Show Context)
Citation Context ...n in Imoto et al. [23]. Note that the proposed prior probabilityof the network can be used for other types of Bayesian network models, such as discrete Bayesian networks and dynamic Bayesian networks =-=[29, 34, 36, 43]-=-. The computation of partition function, �, is intractable even for moderately sized gene networks. To avoid this problem, we compute upper and lower bounds of the partial function and use them for ch... |

33 | A construction of Bayesian networks from databases based on an MDL scheme - Suzuki - 1993 |

23 | Using Protein-protein Interactions for Refining Gene Networks Estimated from Microarray Data by Bayesian Networks
- Nariai, Kim, et al.
- 2004
(Show Context)
Citation Context ...d. If we know gene� and gene� create a protein-protein interaction, we set Ù�� � Ù�� � � .Insuch a case, we will decide whether we make a virtual node corresponding to a protein complex theoretically =-=[35]-=-. Protein-DNA interactions Protein-DNA interactions show gene regulations by transcription factors and can be modeled more easily than protein-protein interactions. When gene� is a transcription regul... |

17 |
Approximate predictive likelihood
- Davison
- 1986
(Show Context)
Citation Context ...terion based on the posterior probability of the network is how to compute the marginal likelihood given by a high dimensional integral. Imoto et al. [23] used the Laplace approximation for integrals =-=[8, 30, 45]-=- and derived a criterion, named BNRC��Ø�ÖÓ (Bayesian network and Nonparametric heteroscedastic Regression Criterion), of the form where �ÆÊ���Ø�ÖÓ � �sÐÓ�� � ÐÓ� ¬ ¬ Ò Ð� ���� � Ò �� Â� �� �s� Â� � ��... |

16 |
Bayesian information criteria and smoothing parameter selection in radial basis function networks
- Konishi, Ando, et al.
- 2004
(Show Context)
Citation Context ...terion based on the posterior probability of the network is how to compute the marginal likelihood given by a high dimensional integral. Imoto et al. [23] used the Laplace approximation for integrals =-=[8, 30, 45]-=- and derived a criterion, named BNRC��Ø�ÖÓ (Bayesian network and Nonparametric heteroscedastic Regression Criterion), of the form where �ÆÊ���Ø�ÖÓ � �sÐÓ�� � ÐÓ� ¬ ¬ Ò Ð� ���� � Ò �� Â� �� �s� Â� � ��... |

12 | A string pattern regression algorithm and its application to pattern discovery in long introns
- Bannai, Inenaga, et al.
- 2002
(Show Context)
Citation Context |

11 | Chain functions and scoring functions in genetic networks - Gat-Viks, Shamir - 2003 |

9 |
Linking microarray data to the literature
- Masys
(Show Context)
Citation Context |

9 | Selection of smoothing parameters in B-spline nonparametric regression models using information criteria - Imoto, Konishi - 2003 |

7 |
Flexible Smoothing with -Splines and Penalties
- Eilers, Marx
- 1996
(Show Context)
Citation Context ...roceedings of the Computational Systems Bioinformatics (CSB’03) 0-7695-2000-6/03 $17.00 © 2003 IEEE and normally on mean 0 and variance � ��. Here, Ñ�� ¡ is a smooth function constructed by �-splines =-=[9, 12, 24]-=- of the form where �� � � Ñ�� Ô � �� � Å�� � Ñ� � � � Ñ��Ñ� Ô�� � � ¡ ������Å���� ¡ � is a prescribed set of �- splines and � Ñ� are parameters. Hence, a Bayesian network and nonparametric heteros... |

4 |
A literature network of human genes for high-throughput analysis of gene expression
- Komorowski, Hovig
(Show Context)
Citation Context ...EGG [28], contain several known gene networks and pathways. This information can be used similarly. Literature Some research has been performed to extract information from a huge amount of literature =-=[27]-=-. Literature contain various kinds of information including biological knowledge described above. So we can model literature information in the same way. 3. Computational Experiments 3.1. Monte Carlo ... |

1 | Selection of smoothing parameters in �-spline nonparametric regression models using information criteria - Imoto, Konishi |

1 |
We do not intend to imply that counter-deception is mainly a process of anomaly detection, in the statistical sense. We use the term “anomaly” to denote evidence that is not consistent with current beliefs about the state of the world or the predicted act
- unknown authors
- 1996
(Show Context)
Citation Context ...thod to Saccharomyces cerevisiae gene expression data in Section 3.2. 2. Method for Estimating Gene Networks 2.1. Bayesian network and nonparametric heteroscedastic regression model Bayesian networks =-=[26]-=- are a type of graphical models for capturing complex relationships among a large amount of random variables by the directed acyclic graph encoding the Markov assumption. In the context of Bayesian ne... |