Integrative bioinformatics: a new approach for the target ... - Aepic

HIGH PERFORMANCE COMPUTING
HIGH PERFORMANCE COMPUTING
FOR INTEGRATIVE
FOR INTEGRATIVE
CARDIOVASCULAR DRUG DESIGN
CARDIOVASCULAR DRUG DESIGN
1
Patrizio Arrigo,
2
Carmelina Ruggiero
1
CNR ISMAC, Genoa, Italy

2
DIST - University of Genoa,
Italy
Changes in Life Science Research in
Changes in Life Science Research in
Pre-genomics and Post-genomic era
Pre-genomics and Post-genomic era
Pre
-
genomics

Post
-
genomics

Small dataset

Massive dataset

Limited data reuse

Extensive data reuse

Limited data sharing

Extensive data
sharing

Straight
-
forward

Complex data
analysis





Data production
and analysis




Manual analysis

Computer bas
ed
analysis

Relative low costs

Very high costs

Hypothesis
-
driven

Data
-
driven



Experimentation

Easy design of
experiments

Complex experiment
design

Reductionist

holistic

Mono
-
disciplinary

Multi
-
disciplinary

experimental

Conce
ptual

Limited collaboration

Extensive
collaborations

Long
-
standing
collaboration

Flexible virtual
organization



Research approach

Conventional science

E
-
science



Molecular medicine approach in the
Molecular medicine approach in the
post-genomics era
post-genomics era
Determination
Determination
Of the genetic
Of the genetic
Component of the
Component of the
Disease
Disease
Gene
Gene
screening
screening
Diagnostics
Diagnostics
screening
screening
Pharmacogenomics
Pharmacogenomics
Gene Therapy
Gene Therapy
Investigation of
Investigation of
Molecular mechanism of
Molecular mechanism of
the diseases
the diseases
Therapeutic
Therapeutic
Target identification
Target identification
And validation
And validation
An Integrative Virtual Drug Design Laboratory
An Integrative Virtual Drug Design Laboratory
(IVDDL)
(IVDDL)
•
Database federation and heterogeneous data sources
integration
•
Development of high performance information and
communication technology drivers that allow to
synchronize and optimize the integrative Knowledge
discovery


process
•
Optimization of the data mining process: computer
intensive tools for specific processes can reduce
computation time
•
Integration with Virtual Screening systems
Requirements:
Requirements:
Basic Architecture for a Virtual Laboratory
Basic Architecture for a Virtual Laboratory
(VL
(VL
)
)
Communication Infrastructure
Communication Infrastructure
Experimental
Experimental
Infrastructure
Infrastructure
(Wet-Laboratory)
(Wet-Laboratory)
Knowledge
Knowledge
Discovery
Discovery
Infrastructure
Infrastructure
Experiment
Experiment
design
design
Wet and
Wet and
In silico
In silico
experiments
experiments
Enhancement
Enhancement
Of Knowledge
Of Knowledge
Decision Support
Decision Support
System Infrastructure
System Infrastructure
Hypothesis
Hypothesis
generation
generation
1
1
2
2
3
3
level
level
CARDIOWORKBENCH Partners
CARDIOWORKBENCH Partners
•
University of Genoa (DIST-Unige), Italy
•
University of Ulster, United Kingdom
•
University of Tartu, Estonia
•
National Research Council (CNR ISMAC), Italy
•
Bielefeld University (UNIBI), Germany
•
University of Leipzig, Germany
•
University of Milan (UMIL), Italy
•
Vienna Medical University (VMU), Austria
•
Fraunhofer Institute for Biomedical Engineering (Fraunhofer), Germany
•
Pharmacelsus (PHARMACEL), Germany
•
Chemilia AB, Sweden
•
Novamass Analytical Ltd (NOVAMASS), Finland
THE CARDIOWORKBENCH PROJECT
THE CARDIOWORKBENCH PROJECT
CARDIOWORKBENCH addresses bottlenecks in the drug design and the optimization
of knowledge transfer in join research activities that involve industrial and academicals
entities. The main activities of the project are the following:
•
DNA Array technology
•
High throughput proteomic analysis
•
Integration of functional genomics and proteomics data mining
•
Experimental dissection of metabolic pathways
•
Determination of structural characteristics of the screened targets
•
Quantitative activity and structural relationship (QSAR) and protein-ligand interaction
analysis
•
Development of ADME and pharmacokinetic analysis system based on “
in silico
”
tissue computational models
•
Evaluation of the interaction among metabolic pathways
•
Modelling cellular behaviour on the basis of metabolic data
•
Optimization of cell culture parameters
CARDIOWORKBENCH Architecture and
CARDIOWORKBENCH Architecture and
workflow
workflow


Mathematical modeling
Of metabolic pathway

Mathematical modeling
Of metabolic pathway

Data Sources used in CARDIOWORKBENCH
Data Sources used in CARDIOWORKBENCH
Gene
Gene
Existing Ligand
Existing Ligand
Database
Database
Phenotype
Phenotype
Genetic varabiality Database
Genetic varabiality Database
DNA microarray
DNA microarray
3D Structure
3D Structure
Proteomic analysis
Proteomic analysis
Interacting proteins
Interacting proteins
Cardiovascular
Cardiovascular
Disease
Disease
Cellular Data
Cellular Data
Metabolic Pathways
Metabolic Pathways
Existing Drug
Existing Drug
Database
Database
Relevant components of proposed GRID
Relevant components of proposed GRID
infrastructure in CARDIOWORKBENCH
infrastructure in CARDIOWORKBENCH
Data Warehouse
Data Warehouse
-Public data (molecular biology and
-Public data (molecular biology and
chemical data)
chemical data)
-Project data (clinical and experimental
-Project data (clinical and experimental
data)
data)
Data Mining
Data Mining
Bioinformatic and chemoinformatic
Bioinformatic and chemoinformatic
Virtual screening
Virtual screening
-Fragment library generation (preliminary to
-Fragment library generation (preliminary to
Docking)
Docking)
-Optimization of structure generation
-Optimization of structure generation
-QSPR/QSAR models development
-QSPR/QSAR models development
CARDIOWORKBENCH functional
CARDIOWORKBENCH functional
Bioinformatics resources
Bioinformatics resources


FUNCTIONAL
FUNCTIONAL
GENOMICS:
GENOMICS:
-Microarray
-Microarray
-RNA Interference
-RNA Interference
PROTEOMICS:
PROTEOMICS:
-
Antibody array
Antibody array
-
Protein chip
Protein chip
Integrative knowledge
Integrative knowledge
discovery
discovery
Chemoinformatics
Chemoinformatics
Knowledge discovery
Knowledge discovery
System biology
System biology
modelling
modelling
CARDIOWORKBENCH
CARDIOWORKBENCH
data base
data base
Chemoinformatic Knowledge discovery
Chemoinformatic Knowledge discovery
Structural property
Structural property
analysis
analysis
Chemical data base
Chemical data base
screening
screening
QSAR analysis
QSAR analysis
Chemical
Chemical
synthesis
synthesis
CARDIOWORKBENCH
CARDIOWORKBENCH
data base
data base
System modeling module
System modeling module
Metabolic compartment modelling
Metabolic compartment modelling
Compartment model integrator
Compartment model integrator
Metabolic compartment modelling
Metabolic compartment modelling
CARDIOWORKBENCH integrated high
CARDIOWORKBENCH integrated high
performance resources
performance resources
DIST Project Server
TARTU Computational
Chemistry Server
BIELEFELD Integrative
Bioinformatic Server
ISMAC system biology
modelling Server
University of Vienna
Data Mining Server
University of Ulster
Advantages of GRID
Advantages of GRID
for CARDIOWORKBENCH
for CARDIOWORKBENCH
•
Distribution of heavy computational tasks
(computational chemistry etc)
•


Agent based technology for information
filtering
•


Flexyble workflow design for integrative
data mining
•
Avalability of the computational tools of
for data communication and exchange