david abramson v2 - australian national university · pdf filedederko, nevo, altshuler, wu,...
TRANSCRIPT
Grid enabling 'real' science and engineering ?
David AbramsonMonash e-Science and Grid Engineering Lab (MESSAGE Lab)
Faculty of Information Technology
Acting Director: Monash e-Research Centre
ARC Professorial Fellow
1
Contents• The changing face of science• Grid testbeds: The Pragma testbed• Software tools for “real” science
– The MESSAGE Lab approach• Some successful projects
– Environmental science, – Systems biology, – Chemistry– Physics– Engineering
• Conclusions 2
THE CHANGING FACE OF SCIENCE
3
4
e-Science• Pre-Internet
– Theorize &/or experiment, aloneor in small teams; publish paper
• Post-Internet– Construct and mine large databases of observational or
simulation data– Develop simulations & analyses– Access specialized devices remotely– Exchange information within
distributed multidisciplinary teams
Source: Ian Foster
“Grids are not just communities of computers, but communities of researchers, of people.”
— Peter Arzberger, UCSD
5
6
Typical Grid Applications• Characteristics
– High Performance Computation– Distributed infrastructure– Instruments are first class resources– Lots of data– Not just bigger – fundamentally different
• Some examples– In-silico biology (See MyGrid)– Earthquake simulation– Virtual observatory– High energy physics– Medical applications– Environmental applications.
COMPUTATIONAL PLATFORMS
7
8
The Grid
• Infrastructure (“middleware” & “services”) for establishing, managing, and evolving multi-organizational federations– Dynamic, autonomous, domain independent– On-demand, ubiquitous access to computing, data,
and services• Mechanisms for creating and managing workflow
within such federations– New capabilities constructed dynamically and
transparently from distributed services– Service-oriented, virtualization
Source: Ian Foster
9
What is a Grid?• Three key criteria
– Coordinates distributed resources …– using standard, open, general-
purpose protocols and interfaces …– to deliver non-trivial qualities of
service.• What is not a Grid?
– A cluster, a network attached storage device, a scientific instrument, a network, etc.
– Each may be an important component of a Grid, but by itself does not constitute a Grid
Source: Ian Foster
10
The (Power) Grid:On-Demand Access to Electricity
Time
Qua
lity,
eco
nom
ies
of s
cale
Source: Ian Foster
11
By analogy, some challenges
Voltage 110 – 220 – 240Frequency 50 – 60 Hz.
THE PRAGMA TESTBED: A GRASS ROOTS GRID
12
PRAGMA Grid TestbedPRAGMA Grid Testbed
32 Clusters from 29 institutions in 14 countries/regions (+ 7 in preparation)
UZurichSwitzerland
NECTECThaiGridThailand
UoHydIndia
MIMOSUSMMalaysia
CUHKHongKong
ASGCNCHCTaiwan
IOIT-HCMVietnam
AISTOsakaUUTsukubaTITechJapan
BIIIHPCNGOSingapore MU
Australia
APACQUTAustralia
KISTIKorea
JLUChina
SDSCUSA
CICESEMexico
UNAMMexico
UCNChile
UChileChile
UMCUSA
UUtahUSA
NCSAUSA BU
USA
ASURCCosta Rica
BESTGridNew Zealand
CNICGUCASChina
AIST
SDSC
NGO
NECTECThaiGrid
7 gfarm sites
ASGCLZUChina
CNIC
Source Cindy Zheng
PRAGMA Testbed
14
What have we learned?• Faults (hardware, software and network) are normal • Testbed supports a rich variety of application needs• Testbed is large,
– resource availability, network bandwidth and contention, distributed resource monitoring, and global name spaces.
• Coordination of resources, monitoring, inter-operable software deployments, and resource/user policies complicated because applications may have different requirements for middleware and resource allocation.
• Users and administrators located across 12 time zones. – challenge in putting together a team with very diverse cultural backgrounds.
• Countries with different levels of network speed, quality, and system resources.
• Strong feedback between application and middleware (e.g. Ninf-G, Nimrod and Gfarm).
Zheng, C., Abramson, D., Arzberger, P., Ayyub, S., Enticott, C., Garic, S., Katz, M., Kwak, J., Lee, B. S., Papadopoulos, P., Phatanapherom, S., Sriprayoonsakul, S., Tanaka, Y., Tanimura, Y., Tatebe, O., Uthayopas, P. “The PRAGMA Testbed: Building a Multi-Application International Grid”, Workshop on Grid Testbeds, 6th IEEE International Symposium on Cluster Computing and the Grid 16-19 May 2006, Singapore. 15
MESSAGE LAB APPROACH
16
17
Supporting the software life-cycle
LowerMiddleware
PlatformInfrastructure
Unix Windows JVM TCP/IP MPI .Net Runtime
Environmental Sciences
Life & Pharmaceutical
Sciences
ApplicationsGeo Sciences
VPN SSH
UpperMiddleware/Tools
Globus GT4
Web Services
Test/Debug
SRB
ExecutionDevelopment Deploy
Deploy Motor
DistANTNimrod/G /O /E Kepler
Nimrod/K
Debug
Worqbench
IE EclipseEclipse
Guard
Eclipse VS NimrodPortal& WS
REMUS
GriddLeSNetFiles
Active Data
ActiveSheets
Excel
Nimrod Development Cycle
18
Prepare Jobs using Portal
Jobs Scheduled Executed Dynamically
Sent to available machines
Results displayed &interpreted
PHYSICS
19
Ionisation Chamber Design Lew Kotler, ARPANSA
20
Front wall
Guard electrode
Collectingelectrode
Ion collection volume(10 mm diameter x21 mm length)
Direction of photons
PolystyreneInsulator
TeflonInsulator
Aluminium
Fieldshapingelectrodes
EGS4 Simulation of response of Aluminium Cavity Chamber to 60Co photons
0.125
0.135
0.145
0.155
0.165
0.175
0 0.1 0.2 0.3 0.4
Thickness of front cap (cm)
Ion
pairs
per
inci
dent
pho
ton
Egs4 calculation Fitted to sum of exponentials
RADIATION SOURCE
PATIENT
IMAGER
Radiotherapy planningGiddy, Chin, Lewis, Welsh e-Science Centre, UK
www.utsouthwestern.edu/.../270177SynergyS.2.bmp
BEAMnrc
EGS
DOSXYZnrc
Outcomes
22
CONVOLUTION /SUPERPOSITION MONTE CARLO
Spezi E 2003 PhD Thesis Med Phys 31(3)
SmartPET - A Compton CameraToby Beveridge, Monash University
23
A SmartPET Detector• Large Volume - 20 x 60 x 60 mm3
• Operating Range 0.1 – 2 MeV• Detector resolution depends on Pulse Shape Analysis
A Compton Camera• Extensive FoV• Multi-resolution Data•Angular precision depends on detector resolution
•Multi-parameter space is difficult to characterise, and optimise, analytically•Monte-Carlo solutions such as GEANT4 are computationally expensive
24
Outcomes
Each pixel (at a particular incident energy) was assigned a separate job
At each point the resolution matrix could be calculated
For a Single Trial242 point-source locations (112 field over 2 orthogonal planes)5 energies (between 140 keV and 1000 keV)2 different detection conditions242 x 5 x 2 x (20 mins per run) = 806 hours
ENVIRONMENT
25
Climate StudiesLynch, Abramson, Görgen, Beringer, Uotila,
Monash University• Extensive savanna eco-systems in
northern Australia• Changing fire regime • Fires lead to abrupt changes in
surface properties– Surface energy budgets– Partititioning of convective fluxes – Increased soil heat flux→ Modified surface-atmosphere
coupling • Sensitivity study: do the fire’s
effects on atmospheric processes lead to changes in highly variable precipitation regime of Australian Monsoon?
• Many potential impacts (e.g. agricultural productivity)
(J. Beringer)
Outcomes
0
10
20
30
40
50
60
70
80
90
100
Mar
08
2006
Mar
12
2006
Mar
17
2006
Mar
22
2006
Mar
26
2006
Mar
31
2006
A pr 0
5 20
06
A pr 0
9 20
06
A pr 1
4 20
06
A pr 1
9 20
06
Apr 2
3 20
06
Apr 2
8 20
06
Ma y
03
2006
Ma y
07
2006
Ma y
12
2006
May
17
2006
Ma y
21
2006
Ma y
26
2006
May
31
2006
Jun
04 2
006
Jun
09 2
006
Jun
14 2
006
Jun
18 2
006
Jun
23 2
006
Jun
28 2
006
Jul 0
2 20
06
Jul 0
7 20
06
Jul 1
2 20
06
Jul 1
6 20
06
Jul 2
1 20
06
Jul 2
6 20
06
Jul 3
0 20
06
Aug
04 2
006
Aug
09 2
006
Aug
13 2
006
Aug
18 2
006
mahar
rocks-52
ume
jupiter
pragma001
amata1
tgc
Total
A Workshop On Earth System Models of Intermediate Complexity28-29 March 2006 at the Bureau of Meteorology Research Centre, Melbourne
SYSTEMS BIOLOGY
28
• Heart disease still leading cause of death
• Understanding the underlyingphysiological mechanisms is cheaper and
faster when experimental studies are performed together with mathematical models & computer simulations
• Studying pathologies• Developing & Testing drugs
Normal Rhythm Ventricular Fibrillation
Cardiac ModellingSher, Gavaghan, Hinch, Noble,
Oxford University
Cardiac Modeling
• Based on experimental data, mathematical models have been developed– ODE’s– Initial conditions– Ion movement in single
cells
Shannon et al. model, 2004
Studying ionic modelsAnna Sher, Oxford
• Examine the effect of various parameters on Ca2+-induced Ca2+ release and on shape of the action potential
• Fit simulated to experimental data • Identify parameter(s) that are critical
to distinguish Ca2+ dynamics within various species
Outcomes• Single cell ionic models allow us to study:
– Whole cell currents during an action potential (AP) – Currents in response to voltage-clamp stimuli– Dynamics of ions such as Ca2+ and Na+ – Force-frequency relationship– etc. Vm [mV]
INCX [A/F]
Blue: controlRed: local NCX 0.1Green: local NCX 0.3
Transient inward current
time (ms)
time (ms)
3.0 Hz APD Restitution
00.020.040.060.080.1
0.120.140.160.180.2
0 0.05 0.1 0.15 0.2
DI (s)
APD
90(s
)
Cardiac ModellingDederko, Nevo, Altshuler, Wu, Mcculloch, Mihaylova,
Kerckhoffs , UCSD
APD Restitution
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0 0.5
DI (s)
APD
90 (s
)
0.5 Hz
1.0 Hz
1.5 Hz
2.0 Hz
3.0 Hz
4.0 Hz
APD Restitution
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0.2
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
DI (s)
AP
D 90
(s)
0.5 Hz
1.0 Hz
1.5 Hz
2 .0 Hz
3.0 Hz
4.0 Hz
1.0 Hz Epicardial Cell Calcium Restitution
00.0005
0.0010.0015
0.0020.0025
0 0.2 0.4 0.6DI (s)
Cal
cium
(a.u
.)
Control
0.01 Iso
0.1 Iso
1.0 Iso
CHEMISTRY
34
Quantum ChemistryWibke Sudholt, Univ Zurich
35
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 0.5 1 1.5 2 2.5 3
Pote
ntia
l U/a
.u.
Radius r/a.u.
A1, A2
B1, B2
C CH
H
H
HH
HFC
H
HH
( ) ( ) ( )222
211eff expexp rBArBArU −+−=
36
Drug docking pipelineBaldridge, Amoreira, Univ Zurich,
Berstis, Kondrick, UCSD
• Goal is to minimize the free binding energy• Use Quantum calculations for more realism
Protein Data Bank
PDB2PQR WHATIF QMView APBS Compute free
binding energy
Add Hydrogen Atoms
Remove the water network
Place ligand
Solve Poisson-Boltzmann equation
ENGINEERING
37
Flame Kernel Growth in Turbulent FlowsTom Dunstan, Karl Jenkins, Cranfield University
38
Laminar Flame solutions
Laminar simulations are carried out for verification of the code and adjustment of the initial reaction parameters.
Fig 1. Non-dimensional flame speed.time step = 3x10-5 s
Fig 2. Progress variable C (top) and Reaction Rate (bottom) at progressive time steps through centre of domain.
NIMROD optimisation results
Mass Fraction Burned describes global state of reaction (from 0 in unburned state to 1 in fully burned).
Fig 3. Mass fraction burned for varying values of pre-exponential factor B* and heat release parameter.
Turbulent Flame propagation
39
From parallel computations using 3843 grids (56 million grid points).
Iso-surface of progress variable, C = 0.5
t = 0.8 ms t = 4.8 ms
t = 9.6 ms t = 14.2 ms
2D slices of progress variable taken at position L/4.
Pre-heat zone (pale blue)
Reaction zone (yellow)
Conclusions
• Significant benefits for science and engineering
• Users want to focus on their research• Tools need to be powerful and hide the
complexity of the underlying platforms• Real studies improve tools• Application studies are significant
component of the research www.csse.monash.edu.au/~davida
40
Grid enabling 'real' science and engineering ?
David AbramsonMonash e-Science and Grid Engineering Lab (MESSAGE Lab)
Faculty of Information Technology
Acting Director: Monash e-Research Centre
ARC Professorial Fellow
1
Contents• The changing face of science• Grid testbeds: The Pragma testbed• Software tools for “real” science
– The MESSAGE Lab approach• Some successful projects
– Environmental science, – Systems biology, – Chemistry– Physics– Engineering
• Conclusions 2
THE CHANGING FACE OF SCIENCE
3
4
e-Science• Pre-Internet
– Theorize &/or experiment, aloneor in small teams; publish paper
• Post-Internet– Construct and mine large databases of observational or
simulation data– Develop simulations & analyses– Access specialized devices remotely– Exchange information within
distributed multidisciplinary teams
Source: Ian Foster
“Grids are not just communities of computers, but communities of researchers, of people.”
— Peter Arzberger, UCSD
5
6
Typical Grid Applications• Characteristics
– High Performance Computation– Distributed infrastructure– Instruments are first class resources– Lots of data– Not just bigger – fundamentally different
• Some examples– In-silico biology (See MyGrid)– Earthquake simulation– Virtual observatory– High energy physics– Medical applications– Environmental applications.
COMPUTATIONAL PLATFORMS
7
8
The Grid
• Infrastructure (“middleware” & “services”) for establishing, managing, and evolving multi-organizational federations– Dynamic, autonomous, domain independent– On-demand, ubiquitous access to computing, data,
and services• Mechanisms for creating and managing workflow
within such federations– New capabilities constructed dynamically and
transparently from distributed services– Service-oriented, virtualization
Source: Ian Foster
9
What is a Grid?• Three key criteria
– Coordinates distributed resources …– using standard, open, general-
purpose protocols and interfaces …– to deliver non-trivial qualities of
service.• What is not a Grid?
– A cluster, a network attached storage device, a scientific instrument, a network, etc.
– Each may be an important component of a Grid, but by itself does not constitute a Grid
Source: Ian Foster
10
The (Power) Grid:On-Demand Access to Electricity
Time
Qua
lity,
eco
nom
ies
of s
cale
Source: Ian Foster
11
By analogy, some challenges
Voltage 110 – 220 – 240Frequency 50 – 60 Hz.
THE PRAGMA TESTBED: A GRASS ROOTS GRID
12
PRAGMA Grid TestbedPRAGMA Grid Testbed
32 Clusters from 29 institutions in 14 countries/regions (+ 7 in preparation)
UZurichSwitzerland
NECTECThaiGridThailand
UoHydIndia
MIMOSUSMMalaysia
CUHKHongKong
ASGCNCHCTaiwan
IOIT-HCMVietnam
AISTOsakaUUTsukubaTITechJapan
BIIIHPCNGOSingapore MU
Australia
APACQUTAustralia
KISTIKorea
JLUChina
SDSCUSA
CICESEMexico
UNAMMexico
UCNChile
UChileChile
UMCUSA
UUtahUSA
NCSAUSA BU
USA
ASURCCosta Rica
BESTGridNew Zealand
CNICGUCASChina
AIST
SDSC
NGO
NECTECThaiGrid
7 gfarm sites
ASGCLZUChina
CNIC
Source Cindy Zheng
PRAGMA Testbed
14
What have we learned?• Faults (hardware, software and network) are normal • Testbed supports a rich variety of application needs• Testbed is large,
– resource availability, network bandwidth and contention, distributed resource monitoring, and global name spaces.
• Coordination of resources, monitoring, inter-operable software deployments, and resource/user policies complicated because applications may have different requirements for middleware and resource allocation.
• Users and administrators located across 12 time zones. – challenge in putting together a team with very diverse cultural backgrounds.
• Countries with different levels of network speed, quality, and system resources.
• Strong feedback between application and middleware (e.g. Ninf-G, Nimrod and Gfarm).
Zheng, C., Abramson, D., Arzberger, P., Ayyub, S., Enticott, C., Garic, S., Katz, M., Kwak, J., Lee, B. S., Papadopoulos, P., Phatanapherom, S., Sriprayoonsakul, S., Tanaka, Y., Tanimura, Y., Tatebe, O., Uthayopas, P. “The PRAGMA Testbed: Building a Multi-Application International Grid”, Workshop on Grid Testbeds, 6th IEEE International Symposium on Cluster Computing and the Grid 16-19 May 2006, Singapore. 15
MESSAGE LAB APPROACH
16
17
Supporting the software life-cycle
LowerMiddleware
PlatformInfrastructure
Unix Windows JVM TCP/IP MPI .Net Runtime
Environmental Sciences
Life & Pharmaceutical
Sciences
ApplicationsGeo Sciences
VPN SSH
UpperMiddleware/Tools
Globus GT4
Web Services
Test/Debug
SRB
ExecutionDevelopment Deploy
Deploy Motor
DistANTNimrod/G /O /E Kepler
Nimrod/K
Debug
Worqbench
IE EclipseEclipse
Guard
Eclipse VS NimrodPortal& WS
REMUS
GriddLeSNetFiles
Active Data
ActiveSheets
Excel
Nimrod Development Cycle
18
Prepare Jobs using Portal
Jobs Scheduled Executed Dynamically
Sent to available machines
Results displayed &interpreted
PHYSICS
19
Ionisation Chamber Design Lew Kotler, ARPANSA
20
Front wall
Guard electrode
Collectingelectrode
Ion collection volume(10 mm diameter x21 mm length)
Direction of photons
PolystyreneInsulator
TeflonInsulator
Aluminium
Fieldshapingelectrodes
EGS4 Simulation of response of Aluminium Cavity Chamber to 60Co photons
0.125
0.135
0.145
0.155
0.165
0.175
0 0.1 0.2 0.3 0.4
Thickness of front cap (cm)
Ion
pairs
per
inci
dent
pho
ton
Egs4 calculation Fitted to sum of exponentials
RADIATION SOURCE
PATIENT
IMAGER
Radiotherapy planningGiddy, Chin, Lewis, Welsh e-Science Centre, UK
www.utsouthwestern.edu/.../270177SynergyS.2.bmp
BEAMnrc
EGS
DOSXYZnrc
Outcomes
22
CONVOLUTION /SUPERPOSITION MONTE CARLO
Spezi E 2003 PhD Thesis Med Phys 31(3)
SmartPET - A Compton CameraToby Beveridge, Monash University
23
A SmartPET Detector• Large Volume - 20 x 60 x 60 mm3
• Operating Range 0.1 – 2 MeV• Detector resolution depends on Pulse Shape Analysis
A Compton Camera• Extensive FoV• Multi-resolution Data•Angular precision depends on detector resolution
•Multi-parameter space is difficult to characterise, and optimise, analytically•Monte-Carlo solutions such as GEANT4 are computationally expensive
24
Outcomes
Each pixel (at a particular incident energy) was assigned a separate job
At each point the resolution matrix could be calculated
For a Single Trial242 point-source locations (112 field over 2 orthogonal planes)5 energies (between 140 keV and 1000 keV)2 different detection conditions242 x 5 x 2 x (20 mins per run) = 806 hours
ENVIRONMENT
25
Climate StudiesLynch, Abramson, Görgen, Beringer, Uotila,
Monash University• Extensive savanna eco-systems in
northern Australia• Changing fire regime • Fires lead to abrupt changes in
surface properties– Surface energy budgets– Partititioning of convective fluxes – Increased soil heat flux→ Modified surface-atmosphere
coupling • Sensitivity study: do the fire’s
effects on atmospheric processes lead to changes in highly variable precipitation regime of Australian Monsoon?
• Many potential impacts (e.g. agricultural productivity)
(J. Beringer)
Outcomes
0
10
20
30
40
50
60
70
80
90
100
Mar
08
2006
Mar
12
2006
Mar
17
2006
Mar
22
2006
Mar
26
2006
Mar
31
2006
A pr 0
5 20
06
A pr 0
9 20
06
A pr 1
4 20
06
A pr 1
9 20
06
Apr 2
3 20
06
Apr 2
8 20
06
Ma y
03
2006
Ma y
07
2006
Ma y
12
2006
May
17
2006
Ma y
21
2006
Ma y
26
2006
May
31
2006
Jun
04 2
006
Jun
09 2
006
Jun
14 2
006
Jun
18 2
006
Jun
23 2
006
Jun
28 2
006
Jul 0
2 20
06
Jul 0
7 20
06
Jul 1
2 20
06
Jul 1
6 20
06
Jul 2
1 20
06
Jul 2
6 20
06
Jul 3
0 20
06
Aug
04 2
006
Aug
09 2
006
Aug
13 2
006
Aug
18 2
006
mahar
rocks-52
ume
jupiter
pragma001
amata1
tgc
Total
A Workshop On Earth System Models of Intermediate Complexity28-29 March 2006 at the Bureau of Meteorology Research Centre, Melbourne
SYSTEMS BIOLOGY
28
• Heart disease still leading cause of death
• Understanding the underlyingphysiological mechanisms is cheaper and
faster when experimental studies are performed together with mathematical models & computer simulations
• Studying pathologies• Developing & Testing drugs
Normal Rhythm Ventricular Fibrillation
Cardiac ModellingSher, Gavaghan, Hinch, Noble,
Oxford University
Cardiac Modeling
• Based on experimental data, mathematical models have been developed– ODE’s– Initial conditions– Ion movement in single
cells
Shannon et al. model, 2004
Studying ionic modelsAnna Sher, Oxford
• Examine the effect of various parameters on Ca2+-induced Ca2+ release and on shape of the action potential
• Fit simulated to experimental data • Identify parameter(s) that are critical
to distinguish Ca2+ dynamics within various species
Outcomes• Single cell ionic models allow us to study:
– Whole cell currents during an action potential (AP) – Currents in response to voltage-clamp stimuli– Dynamics of ions such as Ca2+ and Na+ – Force-frequency relationship– etc. Vm [mV]
INCX [A/F]
Blue: controlRed: local NCX 0.1Green: local NCX 0.3
Transient inward current
time (ms)
time (ms)
3.0 Hz APD Restitution
00.020.040.060.080.1
0.120.140.160.180.2
0 0.05 0.1 0.15 0.2
DI (s)
APD
90(s
)
Cardiac ModellingDederko, Nevo, Altshuler, Wu, Mcculloch, Mihaylova,
Kerckhoffs , UCSD
APD Restitution
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0 0.5
DI (s)
APD
90 (s
)
0.5 Hz
1.0 Hz
1.5 Hz
2.0 Hz
3.0 Hz
4.0 Hz
APD Restitution
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0.2
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
DI (s)
AP
D 90
(s)
0.5 Hz
1.0 Hz
1.5 Hz
2 .0 Hz
3.0 Hz
4.0 Hz
1.0 Hz Epicardial Cell Calcium Restitution
00.0005
0.0010.0015
0.0020.0025
0 0.2 0.4 0.6DI (s)
Cal
cium
(a.u
.)
Control
0.01 Iso
0.1 Iso
1.0 Iso
CHEMISTRY
34
Quantum ChemistryWibke Sudholt, Univ Zurich
35
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 0.5 1 1.5 2 2.5 3
Pote
ntia
l U/a
.u.
Radius r/a.u.
A1, A2
B1, B2
C CH
H
H
HH
HFC
H
HH
( ) ( ) ( )222
211eff expexp rBArBArU −+−=
36
Drug docking pipelineBaldridge, Amoreira, Univ Zurich,
Berstis, Kondrick, UCSD
• Goal is to minimize the free binding energy• Use Quantum calculations for more realism
Protein Data Bank
PDB2PQR WHATIF QMView APBS Compute free
binding energy
Add Hydrogen Atoms
Remove the water network
Place ligand
Solve Poisson-Boltzmann equation
ENGINEERING
37
Flame Kernel Growth in Turbulent FlowsTom Dunstan, Karl Jenkins, Cranfield University
38
Laminar Flame solutions
Laminar simulations are carried out for verification of the code and adjustment of the initial reaction parameters.
Fig 1. Non-dimensional flame speed.time step = 3x10-5 s
Fig 2. Progress variable C (top) and Reaction Rate (bottom) at progressive time steps through centre of domain.
NIMROD optimisation results
Mass Fraction Burned describes global state of reaction (from 0 in unburned state to 1 in fully burned).
Fig 3. Mass fraction burned for varying values of pre-exponential factor B* and heat release parameter.
Turbulent Flame propagation
39
From parallel computations using 3843 grids (56 million grid points).
Iso-surface of progress variable, C = 0.5
t = 0.8 ms t = 4.8 ms
t = 9.6 ms t = 14.2 ms
2D slices of progress variable taken at position L/4.
Pre-heat zone (pale blue)
Reaction zone (yellow)
Conclusions
• Significant benefits for science and engineering
• Users want to focus on their research• Tools need to be powerful and hide the
complexity of the underlying platforms• Real studies improve tools• Application studies are significant
component of the research www.csse.monash.edu.au/~davida
40
1
eResearch Australasia, Brisbane, 27 June 2007
Data Linking and Integrationfor Health Applications
Dr Anthony MaederResearch Director
E-Health Research Centre / CSIRO ICT CentreBrisbane, Queensland, Australia
http://ict.csiro.au/Overview
E-Health and Health Data The HDI Software Tool Current ProjectsFuture Directions
2
http://ict.csiro.au/
E-Health and Health Data
http://ict.csiro.au/Scope of e-Health
Contemporary health care has adopted evidence-based medicine delivered by multi-disciplinary, multi-party health care teams in a patient-centred approach e-Health encompasses the broad application of Information and Communication Technologies, in support of health care needsThe main e-Health domains of activity are:
• Health Information Systems (data and software tools)• Health Services Delivery (work practices and processes)
3
http://ict.csiro.au/Australian Health Systems Scene
State-based systems, with different structures for management of hospitals and community healthNational and local organisation of private providersAustralian Government provides financial resourceMany independent legacy software systems: specialised, non-interoperable, unsupportedSafety, Quality and Efficiency issues driving reform of work practices and wider sharing of informationKey problem is lack of universal health identifier: National e-Health Transition Authority aims at this
http://ict.csiro.au/Need and Benefits of Data Linking
Currently patient data resides across numerous different databases which are unconnected and owned separately
Different information systems and reporting systemsGovernment vs Hospital vs GP vs Allied health systems
Health care improvement opportunities flow from linking this datahigher levels of patient care due to fuller informationextension of evidence-based practicebetter planning or decision making for specific casesimprovements to training and education, safety and quality
4
http://ict.csiro.au/Privacy Issues
National Privacy PrinciplesUse of health data for treatment purposesSecondary use of data for research purposes
State Acts related to Health Data:Health Records and Information Privacy Act (2002) - NSWHealth Records Act (2001) – VictoriaMany others
Organisation principlesOther policies applicable site by siteAccess to data governed by ethics compliance
http://ict.csiro.au/
Source: Victorian Privacy Commissioner
Data Linkage and Integration
5
http://ict.csiro.au/Health Data Sources
National levelMedicare – cost codes identify treatmentsPBS - Pharmaceuticals Benefits Scheme
State levelHealth department hospital admissions dataState based disease-specific data collections Pathology reports & resultsRadiotherapy reportsRadiology reports & imagesRegistries
Hospital levelHospital Information systemHospital pharmaceuticals database
Hospital unitsClinical information systemsUnit specific data sources
Clinical areasClinician based data sources
External sourcesGeneral PractitionersEmergency ServicesAllied health enterprises
http://ict.csiro.au/Data Utilization
Data collection management and organisationKnowledge discovery (population or cohorts) Understanding and comparison of casesPre-processing and reshapingStatistical correlation and analysisData aggregation and integrationIdentify events and trendsHealth awareness and promotion
6
http://ict.csiro.au/Problems to Overcome for Data Linking
Major practical impediments exist for data linkingPatient security and privacy restrictionsDiversity and independence of databasesComplexity of data formats Sophistication of aggregation methods
Existing solutions tend to adopt a heavy approachManual processing to achieve one-off linkingData repositories or warehousingTrusted third party units offering linking servicesFull scale integration and interoperability of all systems
http://ict.csiro.au/
The HDI software tool
7
http://ict.csiro.au/The Health Data Integration Project
Aims to provide novel methods for linking multiple databases, by allowing data custodians to retain control of dataProcesses queries and reporting operations remotely at the data locations (in situ)Allows privacy and security restrictions to be metEnables audit trails of data access operations and users to be producedFully software engineered application product has been produced at EHRC after about 20 person-years of effort
http://ict.csiro.au/The HDI Solution
User Interface
Planner
Authorisation
AuthenticationPEP
HDI DataIntegration
Query Executer
Plan
query
Custodian 1
DataService
Data
DataService
Data
Custodian 2
DataService
Data
query
quer
y
AnalysisReporting
Linking
QueryLinkAnalyseReport
8
http://ict.csiro.au/HDI: Data Custodian Control
Data custodian retains control and securityNo warehousing of dataAll patient-identifying data is encrypted
Databases are added to a HDI installation by the data custodian
The data custodian specifies who can use the data and how they can use it
Metadata layer linked to industry standards to provide a common language and across repositories – increasing usability
http://ict.csiro.au/HDI: Delineation of Responsibility
HDI Domain conceptProvides demarcation of roles and responsibilities
• Domain Administrator
Supports existing ethics committee approval process
• Data Custodian
• Project Administrator
• Project member
9
http://ict.csiro.au/HDI: Building a virtual data collection
Query on linked CRC surgical data and chemotherapy data to get a dataset of information on CRC patients, their current status and their chemotherapy treatments.
CRC Surgical Data
Chemotherapy Record Data
Link Table
Add Search Terms
http://ict.csiro.au/HDI: Building a virtual data collection
Query results show CRC surgical data and chemotherapy data for each de-identified patient.
CRC surgical data
Chemotherapy data
10
http://ict.csiro.au/HDI: Performing an Analysis
Generate the survival chart, for example, Kaplan Meier.Analyse survival outcomes by stage.
Survival chart
Survival analysis statistics
http://ict.csiro.au/HDI: Generating a report
Data from multiple databases can be used for reporting.
Eg… linked data from surgical databases, chemotherapy records and cancer registries
Report on patients given adjuvant chemotherapy
Select report
Select variables on
which to report
Data for report
11
http://ict.csiro.au/
Current Projects
http://ict.csiro.au/Data Linking Projects
Some Current Clinical Applications of HDIQueensland Health: Queensland Oncology On LineRoyal Melbourne Hospital: Colorectal CancerSydney South West Area Health Services: Colorectal CancerAutomated Cancer Staging: Lung Cancer
12
MA
TCH
LI
NK
STA
ND
ARD
ISE
DA
TA L
AY
ER
SO
UR
CE
S
YS
TEM
SC
LIN
ICAL
LA
YE
R
QOOL Application Suite
Queensland Oncology Repository (QOR)
HL7‘Trickle Feed’
Summary of Data Flow
Identifies data sources
Processes the data
Stores the data
Makes the data accessible
Allows users to enter data
on-line
http://ict.csiro.au/Royal Melbourne Hospital: Colorectal Cancer - Overview
ObjectiveIntegration of CRC screening, surgery and family history information:
• For research on sensitivity and specificity of the Faecal Occult Blood Test (FOBT)• For research on surgical outcomes including factors such as adjuvant therapy and
co-morbidities
Cohort Surveillance program patient data over 25 years (approx 3000 records)Surgical data of approx 4000 records; Family history approx 4000 records
Databases3 databases (CRC Screening Program (administrative and outcomes data), Surgical and Clinical CRC data, Family History data)
Data QualitySignificant cleaning work required on databases to regularise recorded data and remove data entry, data format, or historical information change errors
13
http://ict.csiro.au/Royal Melbourne Hospital: Colorectal Cancer - Outcomes
Data cleaning effort and data analysis gave access to 25 years of surveillance information
FOBT sensitivity published
Surgical OutcomesBy factors such as smoking history, adjuvant therapy or diabetes status
SENSITIVITY OF FAECAL OCCULT BLOOD TESTING (FOBT) FOR ASYMPTOMATIC COLORECTAL CANCER AND ADVANCED ADENOMAS OVER A 25-YEAR EXPERIENCE IN COLORECTAL CANCER SCREENING
The performance characteristics of FOBT for asymptomatic colorectal cancer and advanced adenomas is uncertain. Sensitivity, for example, requires direct comparison with a gold standard such as colonoscopy or, less accurately, close follow-up of a cohort of negative screenees to detect interval cancers.
The Royal Melbourne Hospital Bowel Cancer Surveillance Program began in 1979. It coordinates colonoscopicsurveillance (along NHMRC guidelines) and annual FOBT of individuals at moderate and high risk of bowel cancer. Hemoccult-type FOBT was used from 1979 to 1986, immunochemical testing from 1995 to now, and both test types in the interval 1987 to 1994.
A database of FOBT and colonoscopy results has been maintained throughout this time, but the technical resources to examine the data have not been available until now.
These results:
• Show that FOBT identifies a sub-group of at-risk but asymptomatic people enriched with colorectal neoplasia
• Support the need for colonoscopy surveillance in this population
• Illustrate the potential of technologically advanced database interrogation methods such as MMIM
This study also supports introduction of FOBT-based screening for average risk subjects, however, results should be interpreted in the light of our population being at above-average risk
• 4821 registrants have had 5398 colonoscopies since the program started in 1979
• 893 asymptomatic patients with planned colonoscopies had a preceding FOBT within 3 months (see Table)
• 5 cancers were detected, with 3 (60%) being FOBT positive
•There were 49 advanced adenomas, of which 17 (34.7%) were FOBT positive
• There were 148 episodes with detection of any neoplasia, of which 38 (25.7%) were FOBT positive
• Using MMIM, the Surveillance Program database was interrogated to identify asymptomatic patients who had a scheduled colonoscopy and an FOBT within the preceding 3 months
• Patients who had colonoscopy because of a positive FOBT or symptoms were not included
• De-identified data was analysed at the EHRC using HDI (Health Data Integration) and other tools
• The results of the colonoscopy were characterized as normal, carcinoma, advanced adenoma and other adenoma
• ‘Advanced adenoma’ was defined as:
• 3 or more polyps
• size >10mm
• villous component or severe dysplasia
• Sensitivity, specificity and predictive value for neoplasia of FOBT were calculated according to the efficient-score method described by Newcombe
Figure 2 FOBT sensitivity by colonoscopy findings in asymptomatic at-risk people
In collaboration with the EHRC (e-Health Research Centre) and Bio21MMIM (Molecular Medicine Informatics Model) project we reviewed our experience of 25 years of FOBT testing in asymptomatic moderate and high risk subjects undergoing scheduled colonoscopy, to assess the sensitivity, specificity and predictive value for neoplasia of FOBT.
Introduction
Aim
Methods
Results
Conclusion
F.A.Macrae1, M.A.Slattery1, G.J.Brown1, M.A.O’Dwyer2, C.G.Murphy2, M.E. Hibbert3, D.J.St.John1.Colorectal Medicine & Genetics, The Royal Melbourne Hospital1; e-Health Research Centre, CSIRO ICT Centre, Brisbane2; & Bio21
MMIM Project, Melbourne3.
13.419.1
34.7
60
0
10
20
30
40
50
60
70
COLONOSCOPY FINDINGS
Negative
n=745
Other Adenomas
n=94
Advanced Adenomas
n= 49Carcinoma
n=5
FOBT
PO
SITI
VE R
ATE
(%)
FOBTResults
Colonoscopy Findings
Normal OtherAdenoma
Advanced Adenoma
Carcinoma TOTAL
Negative 645 76 32 2 755
Positive 100 18 17 3 138
Total 745 94 49 5 893
Table FOBT results by colonoscopy findings in asymptomatic at-risk people
Figure 1 The immunochemical FOBT kit in current use
• Specificity for advanced adenomas or cancers was 85.9 % (95% Confidence Interval, 83.4% - 88.2%)
• The predictive value for a positive FOBT for any neoplasia in AS subjects was 27.5% (20.4% to 35.9%)
• The likelihood ratio for detection of an advanced adenoma or cancer if the FOBT was positive was 2.6 (1.8 to 3.9)
Results (cont)
SENSITIVITY OF FAECAL OCCULT BLOOD TESTING (FOBT) FOR ASYMPTOMATIC COLORECTAL CANCER AND ADVANCED ADENOMAS OVER A 25-YEAR EXPERIENCE IN COLORECTAL CANCER SCREENING
The performance characteristics of FOBT for asymptomatic colorectal cancer and advanced adenomas is uncertain. Sensitivity, for example, requires direct comparison with a gold standard such as colonoscopy or, less accurately, close follow-up of a cohort of negative screenees to detect interval cancers.
The Royal Melbourne Hospital Bowel Cancer Surveillance Program began in 1979. It coordinates colonoscopicsurveillance (along NHMRC guidelines) and annual FOBT of individuals at moderate and high risk of bowel cancer. Hemoccult-type FOBT was used from 1979 to 1986, immunochemical testing from 1995 to now, and both test types in the interval 1987 to 1994.
A database of FOBT and colonoscopy results has been maintained throughout this time, but the technical resources to examine the data have not been available until now.
These results:
• Show that FOBT identifies a sub-group of at-risk but asymptomatic people enriched with colorectal neoplasia
• Support the need for colonoscopy surveillance in this population
• Illustrate the potential of technologically advanced database interrogation methods such as MMIM
This study also supports introduction of FOBT-based screening for average risk subjects, however, results should be interpreted in the light of our population being at above-average risk
• 4821 registrants have had 5398 colonoscopies since the program started in 1979
• 893 asymptomatic patients with planned colonoscopies had a preceding FOBT within 3 months (see Table)
• 5 cancers were detected, with 3 (60%) being FOBT positive
•There were 49 advanced adenomas, of which 17 (34.7%) were FOBT positive
• There were 148 episodes with detection of any neoplasia, of which 38 (25.7%) were FOBT positive
• Using MMIM, the Surveillance Program database was interrogated to identify asymptomatic patients who had a scheduled colonoscopy and an FOBT within the preceding 3 months
• Patients who had colonoscopy because of a positive FOBT or symptoms were not included
• De-identified data was analysed at the EHRC using HDI (Health Data Integration) and other tools
• The results of the colonoscopy were characterized as normal, carcinoma, advanced adenoma and other adenoma
• ‘Advanced adenoma’ was defined as:
• 3 or more polyps
• size >10mm
• villous component or severe dysplasia
• Sensitivity, specificity and predictive value for neoplasia of FOBT were calculated according to the efficient-score method described by Newcombe
Figure 2 FOBT sensitivity by colonoscopy findings in asymptomatic at-risk people
In collaboration with the EHRC (e-Health Research Centre) and Bio21MMIM (Molecular Medicine Informatics Model) project we reviewed our experience of 25 years of FOBT testing in asymptomatic moderate and high risk subjects undergoing scheduled colonoscopy, to assess the sensitivity, specificity and predictive value for neoplasia of FOBT.
Introduction
Aim
Methods
Results
Conclusion
F.A.Macrae1, M.A.Slattery1, G.J.Brown1, M.A.O’Dwyer2, C.G.Murphy2, M.E. Hibbert3, D.J.St.John1.Colorectal Medicine & Genetics, The Royal Melbourne Hospital1; e-Health Research Centre, CSIRO ICT Centre, Brisbane2; & Bio21
MMIM Project, Melbourne3.
13.419.1
34.7
60
0
10
20
30
40
50
60
70
COLONOSCOPY FINDINGS
Negative
n=745
Other Adenomas
n=94
Advanced Adenomas
n= 49Carcinoma
n=5
FOBT
PO
SITI
VE R
ATE
(%)
FOBTResults
Colonoscopy Findings
Normal OtherAdenoma
Advanced Adenoma
Carcinoma TOTAL
Negative 645 76 32 2 755
Positive 100 18 17 3 138
Total 745 94 49 5 893
FOBTResults
Colonoscopy Findings
Normal OtherAdenoma
Advanced Adenoma
Carcinoma TOTAL
Negative 645 76 32 2 755
Positive 100 18 17 3 138
Total 745 94 49 5 893
FOBTResultsFOBTResults
Colonoscopy FindingsColonoscopy Findings
NormalNormal OtherAdenoma
OtherAdenoma
Advanced AdenomaAdvanced Adenoma
CarcinomaCarcinoma TOTALTOTAL
NegativeNegative 645645 7676 3232 22 755755
PositivePositive 100100 1818 1717 33 138138
TotalTotal 745745 9494 4949 55 893893
Table FOBT results by colonoscopy findings in asymptomatic at-risk people
Figure 1 The immunochemical FOBT kit in current use
• Specificity for advanced adenomas or cancers was 85.9 % (95% Confidence Interval, 83.4% - 88.2%)
• The predictive value for a positive FOBT for any neoplasia in AS subjects was 27.5% (20.4% to 35.9%)
• The likelihood ratio for detection of an advanced adenoma or cancer if the FOBT was positive was 2.6 (1.8 to 3.9)
Results (cont)
F.A.Macrae1, M.A.Slattery1, G.J.Brown1, M.A.O’Dwyer2, C.G.Murphy2, M.E. Hibbert3, D.J.St.John1.Colorectal Medicine & Genetics, The Royal Melbourne Hospital1; e-Health Research Centre, CSIRO ICT Centre, Brisbane2; &
Bio21 MMIM Project, Melbourne3.
http://ict.csiro.au/CRC Services Research Group (NSW) - Overview
ObjectiveDemonstrate the ability to gather patient information across databases and across hospitals to provide summary information on quality and safety of patient care, adherence to clinical guidelines and comparison across hospitalsTo be able to provide this information to clinicians for access to their own data for case management AND for monitoring as above
Cohort Approx 1000 ‘dummy’ patient records with CRC and corresponding administrative and treatment information
DatabasesDifferent databases (Registry, Administrative, Surgical, Chemotherapy) for 3 different hospitals
Data QualityConstructed data included deliberate errors in patient demographics and different formats of clinical information to demonstrate HDI’s ability in inexact matching and transformation of data entries to a selected standard
14
http://ict.csiro.au/CRC Services Research Group (NSW) - Scenario
Hosp A:Admin DB
Hosp B:Admin DB
Hosp C:Admin DB
Hosp A, Surgeon 2:Procedural
Records DB(no chemo data)
Hosp B,Surgeon 3:Procedural
Records DB(no chemo data)
Hosp C, Surgeon 4:
Procedural Records DB
(no chemo data)
Hospital B:Private hospital, not included in Cancer Institute NSW Clinical Cancer Registry Pilot
Hospital C*:Non-teaching hospital, not included in Cancer Institute NSW Clinical Cancer Registry Pilot
Hospital A:Large teaching hospital, included in Cancer Institute NSW Clinical Cancer Registry Pilot
Hosp A, Surgeon 1:Procedural
Records DB(no chemo data)
Hosp A:Cancer Institute
NSW Clinical Cancer Registry
Hosp B,Oncologist 5:
Chemotherapy Records DB
AlphaHosp
CINCCChemoRec
BudgetHosp CHADMIN
CSurgAccSurgAllSurgery BuddsData
*Hospital C – less CRC cases in general, and some harder cases referred to Hospital A or B. All chemotherapy (if required) performed by Hospital A or B.
Hospital B – All chemotherapy (if required) performed by Hospital B.
Hospital A – All chemotherapy (if required) performed by Hospital A.
http://ict.csiro.au/CRC Services Research Group (NSW) - Outcomes
We have demonstrated that with HDI it is possible to report on indicators that require information to be linked within and across hospitals
% o
f P
atie
nts
15
http://ict.csiro.au/Cancer Stage Interpretation System (CSIS)
Staging of a cancer requires access to all available data:
Radiology and histology text reportsInformation extracted from other forms of data, for example, radiological images
Cancer Stage Interpretation System (CSIS)Cancer staging is necessary to determine effective care for individual patients, as well as to design and evaluate health programmes at a population level Develop improved ways to access and analyse stored medical images and reports to better facilitate the staging of cancer patients
http://ict.csiro.au/Cancer Stage Interpretation System (CSIS)
16
http://ict.csiro.au/
Future directions
http://ict.csiro.au/Clinical to Genomic
Integrating biomarker data sources and patient data can provide information on the efficacy, safety and toxicity of drugs
• Advanced non small-cell lung cancer - Iressa effective in only 10% to 15% of patients. Scientists pinpointed mutations in a gene within some tumor cells that allows Iressa to work
Advanced breast cancer - Herceptin issued to 35,000 women world-wide 1998-2005. Only appears to work for patients whose breast cancer cells carry extra copies of the HER2 protein.Micro Array experiments may be useful in finding disease biomarkers
• Linking micro array results with known clinical outcomes will allow new biomarkers to be found
17
http://ict.csiro.au/
Figure by Dr Terry O’Brien, Univ Melbourne (1 July 2005)
Bio21 Molecular Medicine Informatics Model (2005)
http://ict.csiro.au/EHRs and Virtual Registry
Pre-population of Electronic Health RecordUse data linking to gather as much information as possible (to required health record specification) about an individual to pre-populate their complete health recordOutcomes:
• Reduced manual workload for initial set up of health record
Virtual registry Use data linking to gather registry data set information about patients from existing sources (rather than building brand new sources with paper based information submission) – and present back a view of the registry for administration and analysisOutcomes:
• Significantly reduced manual and paper based workload for clinicians and registry officers resulting in more timely registry information, and greater compliance with registry requirements if data capture is at source
18
http://ict.csiro.au/Contemporary ICT Advances
Web ServicesEasy(ier) federation of data and services
OntologiesRelating concepts using semantic propertiesDiscipline based
• Need domain expertise to define the ontology of data sources and fields
The semantic webUsing ontologies in the web applications
http://ict.csiro.au/Using SNOMED CT
Use SNOMED CT for mapping of data to termsStructured data sourcesNatural language
• Reports • Discussions
To make this possibleSubsets for particular domainsAugmenting the domain specific subsetsFast querying
19
http://ict.csiro.au/SNOMED-CT Scope
Clinical TermsComprehensive, not specialty or domain
Human, veterinary, drugs, social, disease, observations, interventions and wellness
~400,000 concepts (fully specified names)~1M descriptions (synonyms etc)~1.4M relationships (900,000 defining)URU principles
Useable, Repeatable, Understandable
http://ict.csiro.au/SNOMED CT Top-level Concepts
138875005 | SNOMED CT Concept123037004 | body structure |404684003 | clinical finding |243796009 | context dependent category |308916002 | environments and geographical locations |272379006 | event |106237007 | linkage concept |363787002 | observable entity |410607006 | organism |373873005 | pharmaceutical / biologic product |78621006 | physical force |260787004 | physical object |71388002 | procedure |362981000 | qualifier value |419891008 | record artifact |48176007 | social context |370115009 | special concept |123038009 | specimen |254291000 | staging and scales |105590001 | substance |
20
http://ict.csiro.au/Expressions, equivalence and subsumption
85189001|acute appendicitissubsumed by• 74400008|appendicitis|• 64572001|disease|:
116676008|associated morphology|=23583003|inflamation|363698007|finding site|=66754008|appendix structure|
equivalent to• 74400008|appendicitis|:
260908002|course|=53737009|acute|
http://ict.csiro.au/Querying
Simple queries can use subsumption/equivalence, for example
find all cases of 74400008|appendicitis| with a particular treatment or outcome
finding those also classified as85189001|acute appendicitisor even as a 64572001|disease| with 23583003|inflamation| of the 66754008|appendix structure|
21
http://ict.csiro.au/
E H E A L T H
http://e-hrc.net
http://ict.csiro.au/
HDI HUB
HDI Data SourceHDI Data Source HDI Data Source
HDI Platform Technology
Custodial controlled
data
HDI integrates
data
De-identified virtuallinked data set
De-identified linked data for
analysisStatistical Packages
e.g. R, SPSS Reporting Toolse.g. Crystal Reports
Custom Applications
Colon Cancer Project
PatientProcedureQuery
ProcedureByEventQuery
Custodial Data
BHospPersonDetails
BHospEventProcedure
BHospProcedure
BHospEventProcedure
BHospProcedure
BHospPersonDetails
DeceasedPersonsQuery
Procedure
EventPersonDetails