Transcript Private - CERN Indico

Plateforme de Calcul pour les Sciences du Vivant
Addressing emerging diseases
on the grid
Vincent Breton, CNRS-IN2P3, LPC Clermont-Ferrand
Credits: Ying-Ta Wu (Academia Sinica, Taïwan)
Doman Kim (Chonnam National University,
Korea)
« Communication is the key to controlling communicable diseases »
Anita Barry, director of Communicable Disease Control, Boston Public Health Commission
http://clrwww.in2p3.fr/PCSV
V. Breton, IFI, 081107
Emerging diseases, a growing burdeon on
public health
Plateforme de Calcul pour les Sciences du Vivant
• Several new diseases have emerged in the last decades (HIV/AIDS,
SRAS, Bird Flu)
• They constitute a growing threat to public health due to world
wide exchanges and circulation of persons
Bird flu status on
January 15th 2008:
- 86 human cases
in 2007, 58 deaths
- 1 lethal case in
2008
- 30 countries
infected by H5N1 in 2007
V. Breton , FCPPL, 150108
Addressing emerging diseases
Plateforme de Calcul pour les Sciences du Vivant
International collaboration is required for:
Prevention (common health policies)
Epidemiological watch
Early detection and warning
Search for new drugs
Search for vaccines
V. Breton , FCPPL, 150108
Searching for new drugs
Plateforme de Calcul pour les Sciences du Vivant
• Drug development is a long (10-12 years) and expensive
(~800 MDollars) process
• In silico drug discovery opens new perspectives to speed it
up and reduce its cost
Target discovery
Lead discovery
Target
Identification and validation
Lead
identification
- 2/5 years
- 30% success rate
- 0.5 year
- 2/4 years
- 65% success rate - 55% success rate
Gene expression analysis,
Target function prediction,
Target structure prediction
De novo design,
Virtual screening
Lead
optimization
Virtual screening,
QSAR
V. Breton , FCPPL, 150108
Screening
Plateforme de Calcul pour les Sciences du Vivant
• Biologists identify a protein
involved in the metabolism of
the virus: the target
• The goal is to find molecules
to prevent the protein from
playing its role in the virus
life cycle: the hits
– Hits dock in the active site of the
protein
• in silico vs in vitro screening
– In silico: computational
evaluation of binding energy
– In vitro: optical measurement of
chemical reaction constant
V. Breton , FCPPL, 150108
Virtual screening workflow
Plateforme de Calcul pour les Sciences du Vivant
Molecular docking
Millions
FLEXX
AUTODOCK
Molecular dynamics
5000
AMBER
Re-ranking
MMPBSA-GBSA
Ligand
Amber
Ligand
Ligand
2 Hydrogen
CHIMERA
4Bonds
H bonds 180
Catalyticaspartic
asparticresidues
residues
Catalytic
Complex
visualization
Catalytic aspartic residues
WET LABORATORY
30
In vitro
tests
Credit: D. Kim
V. Breton , FCPPL, 150108
First large scale grid deployment on
avian flu
Plateforme de Calcul pour les Sciences du Vivant
• Goal n°1: find new drug-like molecules with inhibition
activity on neuraminidase N1, target of the existing
drugs (Tamiflu) against avian flu
– Method: large scale docking of 300.000 selected compounds
against a neuraminidase N1 structure published in PDB
NA
HA
NA is involved in the replication of
virions
Credit: Y-T Wu
V. Breton , FCPPL, 150108
Anticipate the mutations
Plateforme de Calcul pour les Sciences du Vivant
• Emerging diseases
are characterized by
rapidly mutating
viruses
: Predicted mutation site by
structure overlay and
sequence alignment
: Reported mutation site
– Mutations can be
predicted
– Structures can be
modified
• Goal n°2: quantify
the impact of 8
mutations on known
drugs and find new
hits on mutated
targets
V. Breton , FCPPL, 150108
Grid-enabled virtual docking
Plateforme de Calcul pour les Sciences du Vivant
Millions of potential
drugs to test against
interesting proteins!
Compounds:
ZINC: 4.3M
High Throughput Screening
1-10$/compound, several hours
Molecular docking (FlexX, Autodock)
~1 to 15 minutes
Chembridge: 500 000
Targets:
Data challenge on EGEE
~ 2 to 30 days on ~5000 computers
PDB: 3D structures
Cheap and fast!
Selection of the
best hits
Hits screening
using assays
performed on
living cells
Leads
Clinical testing
Drug
V. Breton , FCPPL, 150108
Data challenges on avian flu and
malaria
Plateforme de Calcul pour les Sciences du Vivant
Dates
Target (s)
CPU consumed
EGEE AuverGrid
Data
produced
Specific
features
Status
Summer
2005
Malaria:
plasmepines
80 years
1TB
First data
challenge
In vitro tests
In vivo tests
Spring
2006
Avian flu:
Neuraminidase
N1
100 years*
800 GB*
Only 45 days
needed for
preparation
In vitro tests
Winter
2006
Malaria:
GST, DHFR,
Tubulin
400 years
1,6TB
> 100.000
dockings / hr
Under
analysis
Fall 2007
Avian flu:
Neuraminidase
N1
Estimated 100
CPU years*
Estimated
800 GB*
Joint
deployment
on CNGrid
Data
Challenge
under way
Winter
2007
Malaria:
DHPS
To be estimated
To be
estimated
Joint
deployment
on desktop
grid
In
preparation
*: use of DIANE/GANGA and WISDOM production environments
V. Breton , FCPPL, 150108
Point mutations do impact inhibitory
effectiveness
Plateforme de Calcul pour les Sciences du Vivant
Variation of docking score on wild type
(T06) and mutated targets
Orig.
E119A
E119D
H275F
R293K
E119A_o
T01:E119A
T05:R293K
Y344_o
T01
E119A
V. Breton , FCPPL, 150108
In vitro tests at Chonnam National University
4-Methylumbeliferyl-N-acetyl-a-D-neuramininic acid ammonium salt
[4MU-NANA]; Substrate
First screening
(200 nmol)
Second screening
(2 nmol)
Recombinant Neuraminidase
Spectrofluorometric detector RF-551
362 nm excitation and 448 nm emission wavelengths
Red
Kinetic study
Blue
전남대학
Inhibition
http://altair.chonnam.ac.kr/~carboenz
기능성 탄수화물 효소 및 미생물 유전체 연구실
Results on 308 compounds tested in vitro
4MU-NANA
: 20 mM/RM
Neuraminidase
: 10 mU/reaction
Measure at excitation 362 nm and
emission at 448 nm
Rank
Compounds
Relative activity of
Neu1
1
113
67
2
16
72
3
6
73
4
155
74
5
78
78
63
Tamiflu
100
On UV
전남대학
http://altair.chonnam.ac.kr/~carboenz
기능성 탄수화물 효소 및 미생물 유전체 연구실
The second data challenge
Plateforme de Calcul pour les Sciences du Vivant
• N1 targets
– PDB structures: open and close
conformations (2HU0, 2HU4)
– wild type + 3 mutations (H274, R293,
E119)
– prepared by Italian and Taiwanese
teams (Dr. Luciano Milanesi and Dr.
Ying-Ta Wu)
• Compounds
– 300,000 lab-ready compounds from
Dr. Ying-Ta Wu (Academia SInica,
Taiwan)
– 200,000 compounds from Dr. KunQian Yu (Shanghaï Institute of Materia
Medica, CAS, China)
V. Breton , FCPPL, 150108
Grids for early warning network
Plateforme de Calcul pour les Sciences du Vivant
• Critical importance of
global early warning and
rapid response
– SARS
• Identified keys to set up
successful warning
network
– increased political will
– resources for reporting
– improved coordination and
sharing of information
– raising clinicians'
awareness,
– additional research to
develop more rigorous
triggers for action.
V. Breton , FCPPL, 150108
A data grid to monitor avian flu
Plateforme de Calcul pour les Sciences du Vivant
•
Each database to collect at a national level
– Genomics data on virus and targets
– Epidemiological data: information on human
and bird cases
– Geographical data: maps of outbreaks
– Chemical data: focussed compound libraries
Public
Public
Private
Private
Public
Private
Public
Public
Private
Collaboration started with
Private
IHEP and CNIC within FCPPL:
- Definition of data model
- Implementation using
Public
AMGA
metadata catalogue
Private
V. Breton , FCPPL, 150108
Conclusion
Plateforme de Calcul pour les Sciences du Vivant
•
The grid provides the centuries of CPU cycles required for in silico drug
discovery
– 20% of the compounds selected in silico show better inhibition activity on H5N1
than Tamiflu during in vitro tests
•
The grid offers a collaborative environment for the sharing of data in the
research community on emerging diseases
SCAI Fraunhofer:
Knowledge extraction,
Chemoinformatics
LPC Clermont-Ferrand:
Biomedical grid
CEA, Acamba project:
Biological targets,
Chemogenomics
HealthGrid:
Biomedical grid,
Dissemination
Univ. Los Andes:
Biological targets,
Malaria biology
Univ. Modena:
Biological targets,
Molecular Dynamics
ITB CNR:
Bioinformatics,
Molecular modelling
Univ. Pretoria / CSIR:
Bioinformatics, Malaria
biology
Chonnam nat. univ.:
In vitro testing
KISTI:
Grid technology
Academica Sinica:
Grid user interface
Biological targets
In vitro testing
Mahidol Univ.:
Biochemistry,
in vitro150108
testing
V. Breton , FCPPL,
17
Perspectives
Plateforme de Calcul pour les Sciences du Vivant
• Avian flu
– In vitro tests of the compounds selected in silico for mutated targets
– Second data challenge under way to be analyzed in Taïwan
– Set-up of data repositories with grid data management services
• Other diseases
– Malaria
 already 2 compounds identified with strong inhibition activity on the parasite
-> patent
 In vitro tests planned for in silico selected compounds on 2 targets docked
in the winter of 2006
 New target ready to be deployed both on EGEE and Africa@home
– Diabetes
 Large scale docking started 2 days ago on amylase (CNU, KISTI, LPC)
– AIDS
 Collaboration between Univ. Cyprus and ITB-CNR
V. Breton , FCPPL, 150108
Credits
Plateforme de Calcul pour les Sciences du Vivant
• Development of the WISDOM environment
–
–
–
–
ASGC: Yu-Hsuan Chen, Li-Yung Ho, Hurng-Chun Lee
ITB-CNR: G. Trombetti
CNRS-IN2P3: V. Bloch, M. Diarena, J. Salzemann
HealthGrid: B. Grenier, N. Spalinger, N. Verhaeghe
• Biochemical preparation and analysis
–
–
–
–
ASGC: Y-T Wu
Chonnam National University: D. Kim & al
CNRS-IN2P3: A. Da Costa, V. Kasam
ITB-CNR: L. Milanesi & al
• Projects supporting WISDOM
– Projects providing human resources: BioinfoGRID, EGEE, Embrace
– Projects providing computing resources: AuverGRID, EELA, EGEE,
EUMedGRID, EUChinaGRID, TWGrid
V. Breton , FCPPL, 150108