Continuous Dynamic Bayesian

hartebeestgrassΤεχνίτη Νοημοσύνη και Ρομποτική

7 Νοε 2013 (πριν από 3 χρόνια και 11 μήνες)

72 εμφανίσεις

Continuous Dynamic Bayesian
Network For Gene Regulatory
Network Modeling

Presented by:
Norhaini

Binti

Baba

Supervisor: Dr.
Mohd

Saberi

Bin
Mohamad


Outline


Introduction


Literature
review

Research
Methodology

Materials &
Methods

Conclusion



Introduction

Literature
review

Conclusion

Materials and
methods

Research
methodology

Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline




Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

INTRODUCTION

Gene regulatory network


Collection of DNA segment


Plays role in governing the rates at which gene are transcribed
into mRNA to protein.



Introduction


Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline




Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

Continuous Dynamic Bayesian Network


Directed graphical models of stochastic process


Contain nodes


gene, edges


interaction between genes


Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline




Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

Problem background

Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline




Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

Parametric model provide
limited information on
GRNs

(
Tarmo

Aijo

et al
., 2009)

Difficult of handling cyclic
network of gene regulation

(
Sunyong

Kim
et al
., 2004)

Discretisation might cause
information loss

(
Sunyong

Kim et al
., 2003)

Objectives

Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline




Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

To study and understand the framework of the continuous Dynamic
Bayesian Networks.


To construct gene regulatory networks and evaluate the
performance based on implementation of continuous Dynamic
Bayesian Networks.


To model the continuous Dynamic Bayesian Network for the
construction of gene regulatory networks using the
Bayes

Net
Toolbox.


Scope


Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline




Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

Justification


Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline




Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

Can optimize the network structure, which can give
best representation of the gene interactions

Can analyze the data as a continuous data

Avoiding the ambiguity about edge directions common
to static Bayesian network

Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline




Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

LITERATURE REVIEW

Gene regulation


Process of cell or viruses regulate gene into gene product


Occur at 4 level in eukaryotes.


Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline




Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

Gene Regulatory Network


Genes


nodes.


The interaction between any pair of genes


edges.


Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline




Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

Microarray technology


To monitor and quantify changes in gene expression.


Lead to revolution in speed and scope of genetic analysis of
regulatory networks.


Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline




Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

Gene Expression Data


Structure of how data is organized after hybridization of DNA
microarray.



Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline




Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

Comparison of Previous Methods


Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline




Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline




Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

RESEARCH METHODOLOGY

Research framework


Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline




Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

Phase
1


The studying and understanding the framework of
dynamic Bayesian network.

Phase
2


The modeling of dynamic Bayesian network using
BNT.

Phase
3


The construction of gene regulatory network and
evaluation of performance.

Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline




Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

Data and
resource
collection

Imputation
missing value

Model cDBN using
BNT

Visualize DAG (
graphviz
)

Compare with KEGG database

Find optimal network give best
performance

Data sources


Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline




Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling


Saccharomyces

cerevisiae

Dataset



Escherichia coli

Dataset


Spellman et al.,
1988

DNA Microarray
Datasets website

Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline




Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

MATERIALS & METHODS

Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline




Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

Imputation missing value


Prepare a complete dataset for constructing dynamic Bayesian
networks.


Missing values must be imputed.


Using
knnimpute




Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline




Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

>> A = data;

>>
ecoli

=
knnimpute

(A)

Algorithm to model cDBN in BNT


Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline




Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

Learning


Parameter learning

Structure learning


-
Parameter estimation for continuous
variables.

-
Maximum Likelihood estimation.

-
Data


䍐Cs




-
Constrained
-
based : test independencies

-
Search
-
and
-
score: define selection of
criterion that measures goodness of a
model based on scoring function.

-

Select the highest scoring model.



Construct gene regulatory network


Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline




Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

Implement
the cDBN model in
BNT using
Saccharomyces

cerevisiae

and
Escherichia coli

datasets.



Gene regulatory constructed is
displayed in DAG using
GraphViz
.

Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline




Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

CONCLUSION

Contributions


Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline




Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

cDBN

-

to construct the
gene regulatory network

Overcome the problem
of
discretisation

Identify the best
performance of gene
regulatory network

Future works


Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline




Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

Model the continuous
Dynamic Bayesian Network
using the
Bayes

Net Toolbox

Construct gene regulatory
network and visualize the
DAG using
GraphViz


Compare gene network with
KEGG database to evaluate
the performance

Conclusion



Continuous dynamic
bayesian

network constructed
systematically and criteria for learning network structures for
continuous data.



Advantage of using continuous dynamic
B
ayesian network is
discretisation

is not required.



Helps to study interaction of gene regulation.






Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline




Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline


Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

Q & A

Literature
review

Research
Methodology

Materials &
Methods

Conclusion

Outline


Introduction


Continuous Dynamic Bayesian Network For Gene Regulatory Network Modeling

THANK YOU