The Role of Multiscale Modelling in
Systems Medicine
REPORT
CASyM strategic modelling workshop
11 June 2013
Heidelberg, Germany
CASyM strategic modeling workshop:
The Role of Multiscale Modelling in Systems Medicine
IMPRINT
Publisher
CASyM administrative office
Project Management Jülich, Forschungszentrum Jülich GmbH
Authors
Authors of the report: Olaf Wolkenhauer, Charles Auffray, Andreas Deutsch, Dirk Drasdo, Franceso
Gervasio, Luigi Preziosi, Philip Maini, Anna Marciniak-Czochra, Christina Kossow, Lars Küpfer, Katja
Rateitschak, Ignacio Ramis-Conde, Benjamin Ribba, Andreas Schuppert, Rod Smallwood, Georgios
Stamatakos, Felix Winter, Helen Byrne.
Pictures
O. Wolkenhauer
Date
19 July 2013
Please take note that the content of this document is property of the CASyM consortium. If you wish to
use some of its written content, make reference to: CASyM report: The Role for Multiscale Modelling in
Systems Medicine, Heidelberg, June 2013.
2
CASyM strategic modeling workshop:
The Role of Multiscale Modelling in Systems Medicine
TABLE OF CONTENTS
1 THE ROLE OF MULTISCALE MODELLING IN SYSTEMS MEDICINE ............................................... 5
1.1 The State-of-the-Art in Modelling Biological Systems ......................................................... 6
1.2 Multiscale Modelling ............................................................................................................ 7
1.3 Challenges and Opportunities .............................................................................................. 8
2 WORKSHOP OUTCOMES ............................................................................................................. 9
2.1 Priorities identified during the workshop ............................................................................ 9
2.2 Short-term recommended actions....................................................................................... 9
2.3 Medium-term recommended actions.................................................................................. 9
2.4 Longer-term recommended actions .................................................................................. 10
3 LIST OF PARTICIPANTS............................................................................................................... 11
4 ACKNOWLEDGEMENTS ............................................................................................................. 12
3
CASyM strategic modeling workshop:
The Role for Multiscale Modelling in Systems Medicine
Priority
Figure 1:issues.
Priority issues (source: O. Wolkenhauer)
4
CASyM strategic modeling workshop:
The Role for Multiscale Modelling in Systems Medicine
1
THE ROLE OF MULTISCALE MODELLING IN SYSTEMS MEDICINE
The (mal)functioning of the human body is a complex process, characterised by multiple interactions
between systems that act across multiple levels of structural and functional organisation -- from
molecular reactions to cell-cell interactions in tissues to the physiology of organs and organ systems.
Over the last decade, we have gained detailed insights into the structure and function of molecular,
cellular and organ-level systems, with technologies playing an important role in the generation of data
at these different scales. An important (and challenging) theme for Systems Medicine is the integration
of this knowledge across the relevant levels of organisation.
The human body can be characterised as a multilevel
organisation. For Systems Medicine, the levels of structural
organisation include the molecular, cellular, tissue, organ
and whole organism levels. Structural levels are closely
linked to spatial scales. The functional organisation of the
human body describes the behaviour of (sub)systems and
their role (e.g. cell functions, like apoptosis and
differentiation, in the context of tissue remodelling). The
processes involved in functional organisation typically act
over markedly different temporal scales, ranging from protein
folding to biochemical reactions, the physiology of an organ
and aging. When considering functional organisation at the
cell level one can distinguish three classes of processes:
gene regulation, metabolism and signalling. The
investigation of cell functions is frequently associated with
specific technologies for data generation (e.g. microarrays
for gene expression, mass spectrometry for studying
metabolic networks, or immunoblotting for signalling
pathways). For Systems Medicine these organisational
levels and data need to be integrated. This requires
methodologies that can conceptually and computationally
link or integrate across spatial and temporal scales. Several
different approaches may be used to distinguish the different
levels of abstraction involved in a particular process, say
coarse-grained, rule-based or logical representations, or
detailed mechanistic models that account for biochemical
and biophysical details. More recently, experimental and
modelling techniques have been advanced to analyse tissue
and organ architecture and function as well as their links to
whole body physiology. In summary, multi-levelness and
integration of knowledge are two key concepts in Systems
Medicine that can be addressed through multiscale
modelling.
Mathematicians,
physicists,
engineers and computer scientists, working
with biologists, pharmacists and clinicians,
have developed a range of methodologies to
represent complex nonlinear dynamical
systems. Under the umbrella of systems
biology, workflows of data-driven modelling
and model-driven experimentation have led
to the development of computational
models that describe processes at all levels,
spanning gene regulatory networks, signal
transduction pathways and metabolic
networks, cell populations, structured
tissues as well as pharmacokinetic and
pharmacodynamic (PK/PD) models that
analyse drug action at the whole organism
level, and pharmacogenomic models of
disease risk and drug exposure at the patient
population level.
One of the key challenges facing
systems medicine is the integration of data,
models and knowledge arising from these
efforts, with the goal being to generate a
comprehensive mechanistic understanding
of the (mal)functioning of tissues and organs,
and the therapeutic effect of drugs as well
occurrence of off-target events in order to
develop advanced diagnostic and prognostic
tools and optimised treatments at reduced
number of animal experiments.
In a workshop, funded and organised through the FP7 coordinating and support action CASyM –
Implementation of Systems Medicine across Europe, the role of computational multi-scale models for
Systems Medicine was discussed.
5
CASyM strategic modeling workshop:
The Role for Multiscale Modelling in Systems Medicine
A number of clinically relevant questions formed the background and motivation for this meeting, and
were predicated on the need for a better understanding of:
 the origins of variability in the response of patients to drugs, including intrinsic factors and
extrinsic stimuli from the environment;
 the mechanisms that link the biochemistry of molecular interactions, with biophysical processes
associated with the interactions of cells with each other and their micro-environment;
 the relationships between tissue material properties, tissue architecture and tissue function;
 the ways in which clinical and patient data can be analysed in combination with basic research
data from pre-clinical experimental models.
While large and long-term research initiatives, like the Virtual Physiological Human, the Virtual Liver or
the Human Brain Project, are aiming to develop comprehensive, computational representations of
organs and organ systems, the CASyM workshop focussed on opportunities for smaller, interdisciplinary
collaborations between clinicians and modellers targeting specific questions of clinical relevance.
A major hurdle for the development of quantitative and predictive models that span the gap
between cell-level biochemical models and organism-level
PK/PD models, is the technical difficulty associated with
generating sufficiently comprehensive quantitative
datasets for large numbers of system variables, across
different levels of organisation. To handle the variety of
available data, the development of efficient tools for the
identification of relevant data subsets from existing
repositories is a major challenge for efficient modelling.
Ultimately, this will lead to the development of models
that can be used in clinical practice while accounting for
the uncertainty in the data.
Rapid advances in measurement technologies
(e.g. next generation sequencing) and the apparent Discussing priority issues.
"flood" of data could potentially lead to an overemphasis
of data management issues. While there is no doubt a need for advanced IT infrastructures, it is
important to realise that data will only be useful if interpreted. Mathematical modelling provides a
conceptual framework to guide the interpretation of data, generated from complex biological systems.
In projects envisaged by the workshop participants, modelling is not the final goal; rather it is a tool that
can be used to advance understanding of the system, to develop more directed experiments in the
laboratory, and, ultimately, to generate testable predictions enabling improved therapies. More
important than numerical predictions however, is the process of modelling as a way of thinking about
complex systems, to critically assess the relevant system variables and their interactions, to reveal the
dominant biological processes underlying observed dynamics, to limit outcomes to plausible ranges and
to highlight uncertainties (Epstein 2008).
1.1
The State-of-the-Art in Modelling Biological Systems
The fields of theoretical and mathematical biology have pioneered the development of mathematical
and computational models of biological systems. Systems biology has contributed workflows for datadriven modelling and model-driven experimentation to the life sciences. Taken together this provides a
considerable body of experience for modelling at different levels:
6
CASyM strategic modeling workshop:
The Role for Multiscale Modelling in Systems Medicine
(1) Molecular modelling, e.g.:
 Drug target interactions
 Prediction of pathogenic mutations
 Protein-protein interactions
(2) Modelling of subcellular processes, e.g.:
 Signalling pathways, gene regulatory networks, metabolic networks
 Drug target prediction by sensitivity analysis
 Linking signalling pathways to phenotypic changes, e.g. regression models
(3) Individual- or cell-based models, e.g.:
 Individual behaviour of cells in their microenvironment
 Cell-cell interactions in heterogeneous populations
 Control of cell material properties and shape by molecular mechanisms and the effect of
physical properties of cells and their environment on cell decisions
(4) Tissue/organ level models, e.g.:
 Population dynamics, the formation and maintenance of tissue architecture
 Cell-stroma interactions, e.g. EMT
 Mechano-biology / biomechanical models
(5) Body systems level models, e.g.:
 Pharmacokinetic/pharmacodynamic (PKPD) modelling
 Physiologically based PK/PD (PB-PK/PD) modelling for applications in systems pharacology
 Modelling environmental influences.
The list above is necessarily incomplete and serves here to indicate the progression of modelling efforts,
with an increasing diversity of approaches available, a tendency to combine approaches in modelling
workflows and clear indications for an increasing number of successful case studies. Key challenges that
emerge in Systems Medicine are (i) how to integrate experimental data from a wide range of
technologies and sources, (ii) how to relate different modelling formalisms and (iii) how to integrate
computational implementations. This challenge can be addressed through multiscale modelling.
1.2
Multiscale Modelling
Complex clinical conditions are characterised by defects acting at more than one structural and
functional level. Their multiscale nature represents an important challenge for experimentalists,
clinicians and modellers alike. Multiscale modelling provides a conceptual framework for the efficient
integration of information and data from genomic, cellular, tissue and organ levels up to the whole body
scale. The aim is to enhance clinical decision-making and ultimately to allow prognostic predictions,
improved diagnosis and optimised therapies for individual patients.
Mathematical models serve as a "macroscope" that integrates evidence available at different
levels of a particular system’s structural and functional organisation while, at the same time, enabling us
to zoom in on the details where necessary. For experimental models almost certainly not all parameters
influencing a disease will be known and frequently known parameters cannot be influenced. This makes
it difficult to unambiguously identify the mechanisms at work in diseases. Mathematical modelling
makes an epistemic contribution by allowing "in silico" (simulation) experiments under defined
7
CASyM strategic modeling workshop:
The Role for Multiscale Modelling in Systems Medicine
conditions, aiding the exploration of hypotheses by explicitly accounting for the uncertainty that arises
from biological complexity.
In the context of Systems Medicine, the challenge that needs to be addressed through focussed
research programmes is the development of computationally efficient and clinically validated models
that can support clinicians in the targeted design of individualized therapies with optimal risk-benefit
ratios.
Workflows to develop computational models provide a conceptual framework for formulating
coherent hypotheses derived from patient data (clinical data) and basic research (pre-clinical data). Such
workflows can be used to create models that assimilate evidence at all levels and from a wide range of
data sources. These models are then the basis for the design of experimental studies and simulations to
create and validate focussed predictive models. The actual type of the model should be chosen
according to the requirements and a priori knowledge for the respective project. In practice it can be
mechanistic (if full understanding is available), data-driven or mixed approaches.
1.3
Challenges and Opportunities
With respect to clinical challenges and opportunities related to multiscale modelling, the workshop
participants highlighted the following:











The identification of relevant data and extraction of the model-specific information.
The integration of data and information across levels and from a wide range of sources
(technologies, patients, qualitative and quantitative, omics data, life style data…).
The identification of relevant data and extraction of the model-specific information.
The application of mathematical models to inform clinical decision making; diagnostic tools.
Providing an understanding of mechanisms underlying normal and pathological processes.
Ultimately to allow predictions about disease development (prognosis), drug effects, side effects,
quality of life, dosing schedules and individualized/personalized therapies.
Embedded models in therapeutic devices.
Optimisation of therapeutic interventions as well as risk-benefit ratios for individual.
Continuous knowledge integration along the drug development process
Facilitate demonstration, doctor-patient communication.
Support regulatory decision making during drug approval.
For the more theoretical research problems, the following points should receive attention:




Integration of different levels of abstraction and different model formalisms, e.g the integration of
ODEs with regression and machine-learning based models or for instance bone mechanics and gene
regulation, integrating PDEs for cell density and agent-based models for the cell population.
› Integration of different time scales.
› Representation of different spatial scales.
Concepts, standards and methodologies for the construction, validation and comparison of models,
and to enable re-use of models in different contexts.
Methods to provide computational realisations with reduced computational costs.
The development of verification and testing procedures for complex multiscale models embedded
in clinical support tools is essential for ensuring patient safety.
8
CASyM strategic modeling workshop:
The Role for Multiscale Modelling in Systems Medicine
A particularly important challenge is to enable clinical validation of models and simulations. To this end,
a better understanding and translation from in vitro and animal models to patients is important.
Towards this end, there is a need to bring the multiscale modelling community to work closely together
with clinicians in jointly defined projects.
2
WORKSHOP OUTCOMES
2.1
Priorities identified during the workshop
The workshop identified the following as the main priority for multiscale modelling on a European
Systems Medicine roadmap:
Computational models that help to integrate data and knowledge from the clinics and basic
science (in vitro and animal model experiments) and are applicable to individual patients,
aiming at a mechanistic understanding of pathologies or support of therapy optimisation. This
requires actions for fhe development of concepts, methods and tools that support the
integration of organisational levels to develop interfaces between different computational and
mathematical frameworks used in systems medicine.
2.2
Short-term recommended actions
With respect to data collection and modelling, in the short-term (ie the next one to two years), the
following actions should be given priority:





Exploitation of existing data as a starting point for multiscale modelling, leading to the identification
of gaps in data and in understanding of underlying mechanisms in order to improve the targeted
generation of new data that can be exploited for quantitative models.
To develop SOPs and quality standards for the systematic collection of quantitative data and
information enabling models in a focussed application to be built and validated.
Define test scenarios and proof of concept studies.
Identify required standards and ontologies (e.g. a markup-language and ontology for individualbased models) for models and data repositories in Systems Medicine.
Develop concepts for dedicated modelling workflows for the integration of data and models.
These research actions would benefit from initiatives that (i) generate opportunities to coordinate the
training and joint collaborations between clinicians and multiscale modellers; (ii) support coordinated
clinical projects to bring clinicians, modellers and biologists together; (iii) run sandpits for modellers and
clinicians to develop proposals to address clinical questions, to improve mutual understanding and
realistic expectations for such collaborations.
2.3
Medium-term recommended actions
In the medium-term of two to three years, attention should focus on the provision of suitable IT
infrastructure and the development of standards:


The development of computational tools and algorithms for efficient multiscale simulations.
The development of mathematical formalism to analyse and compare multiscale models such as
parameter estimation, sensitivity analysis, identifiability analysis, image analysis.
9
CASyM strategic modeling workshop:
The Role for Multiscale Modelling in Systems Medicine
Support the development of workflows for modelling, including computational tools that support
data management, model construction and analysis.
 Methods to integrate different physical phenomena, including those of electrical, mechanical and
chemical origin.
 Methods to investigate the interplay between the environment, cell behaviour and cell fate.
These research actions would benefit from initiatives that (i) encourage full-time collaborations between
clinicians and modellers, and allow modellers to spend time in clinics. (ii) IMI-like schemes whereby
clinical groups name prioritized research questions and the consortia form to address these. In the
medium term, the goal is to develop multiscale models of normal physiology and disease, their
development being driven by clinical questions and providing the basis for clinical validation in the
longer term. This will require technologies that promote and facilitate the collaboration of
multidisciplinary teams.

2.4
Longer-term recommended actions
For a time-horizon beyond four years, the focus should be on the application and validation of
multiscale models in the clinic. To this end, it is desirable to:


Enhance the formation of small-scale networks focussed on specific clinical needs, possibly
clustered in a larger integrated project (longer time scale).
Have in place a funding model for small groups of 2-3 partners (1-2 modelling postdocs, 1-2
experimental postdocs + consumables, travel between labs).
Multiscale modelling of biological systems is a cornerstone of Systems Medicine. Multiscale modelling
is the key tool for the integration of data and knowledge, providing the ‘systems’ element of Systems
Medicine. Multiscale modelling therefore lies at the centre of the long-term vision for Systems
Medicine.
Round table discussion.
10
CASyM strategic modeling workshop:
The Role for Multiscale Modelling in Systems Medicine
3
LIST OF PARTICIPANTS
PARTICIPANTS
NAME
FIRST NAME LEGAL ENTITY; CONTACT
Byrne
Helen
Deutsch
Andreas
Drasdo
Dirk
INRIA Paris Rocquencourt, France; [email protected]
Gervasio
Francesco
UCL, UK; [email protected]
Kossow
Christina
Maini
Philip
MarciniakCzochra
Anna
Preziosi
Luigi
Ramis-Conde
Ignacio
Rateitschak
Katja
Ribba
Benjamin
Smallwood
Rod
Stamatakos
Georgios
Winter
Felix
Wolkenhauer
Olaf
School of Mathematical Sciences and Department of Computer
Science, University of Oxford, UK; [email protected]
Dept. of Computer Science, Technical University Dresden;
[email protected]
Dept. of Systems Biology & Bioinformatics, University of Rostock,
Germany; [email protected]
Centre for Mathematical Biology, University of Oxford, UK;
[email protected]
University of Heidelberg, Germany; [email protected]
Dept. of Mathematical Sciences, Politecnico di Torino, Italy;
[email protected]
The Institute of Applied Mathematics in Science and Engineering
(IMACI), Spain; [email protected]
Dept. of Systems Biology & Bioinformatics, University of Rostock,
Germany; [email protected]
INRIA Grenoble, France; [email protected]
Kroto Research Institute, University of Sheffield, UK;
[email protected]
Institute of Communication and Computer Systems,
National Technical University of Athens, Greece;
[email protected]
Dept. of Systems Biology & Bioinformatics, University of Rostock,
Germany; [email protected]
Dept. of Systems Biology & Bioinformatics, University of Rostock,
Germany; [email protected]
11
CASyM strategic modeling workshop:
The Role for Multiscale Modelling in Systems Medicine
4
ACKNOWLEDGEMENTS
The strategic modeling workshop “The Role for Multiscale Modelling in Systems Medicine” is part of the
CASyM work package 3 – “The technological and methodological basis of Systems Medicine”.
CASyM is funded by the European Union, Seventh Framework Programme under the Health Cooperation
Theme and Grant Agreement # 305033.
STEERING COMMITTEE
The following officials, as part of the Scientific Steering Committee, are involved in the scientific
coordination of CASyM:
Charles Auffray - European Institute for Systems Biology & Medicine - EISBM, France
Mikael Benson (Deputy Speaker) - Linköping University Hospital, Sweden
Rob Diemel - The Netherlands Organisation for Health Research and Development, The Netherlands
David Harrison (Speaker) - University of St. Andrews, United Kingdom
Walter Kolch - University College Dublin, Ireland
Frank Laplace - Federal Ministry of Education and Research, Germany
Francis Lévi - Institut National de la Sante et de la Recherche Medicale, France
Damjana Rozman (Deputy Speaker) - University of Ljubljana, Faculty of Medicine, Slovenia
Johannes Schuchhardt - MicroDiscovery GmbH, Germany
Olaf Wolkenhauer - Dept. of Systems Biology & Bioinformatics University of Rostock, Germany
ADMINISTRATIVE OFFICE AND COORDINATION
Marc Kirschner - Project Management Jülich, Forschungszentrum Jülich GmbH, Germany
ORGANIZING COMMITTEE
Marc Kirschner - Project Management Jülich, Forschungszentrum Jülich GmbH, Germany
Christina Kossow - Dept. of Systems Biology & Bioinformatics, University of Rostock, Germany
Katja Rateitschak - Dept. of Systems Biology & Bioinformatics, University of Rostock, Germany
Felix Winter - Dept. of Systems Biology & Bioinformatics, University of Rostock, Germany
Olaf Wolkenhauer - Dept. of Systems Biology & Bioinformatics, University of Rostock, Germany
We are grateful to Prof Roland Eils and Ulrike Conrad from BioQuant Heidelberg for the provision of a
meeting room. Christina Kossow and Felix Winter summarised the discussions and thereby helped to
provide the information on the basis of which the report was written.
The text of the workshop report was written and agreed upon by all participants of the workshop. The
following persons were not present at the workshop but were involved in the preparations and/or
writing of the report: Charles Auffray, EISBM, France; Lars Kuepfer and Andreas Schuppert, Bayer
Technology Services GmbH and RWTH Aachen, Germany.
12
CASyM strategic modeling workshop:
The Role for Multiscale Modelling in Systems Medicine
5
BIBLIOGRAPHY
Epstein JM (2008): Why Model? Journal of Artificial Societies and Social Simulation. Vol 11, No 4 12.
□
13
CASyM strategic modeling workshop:
The Role for Multiscale Modelling in Systems Medicine
14