Developing a Framework for Multi-Scale Modeling of the Digital

Patient: Insights from Current Status and Future Directions

C. Donald Combs

, Lubna Pinky, Chathurani Ranathunge, Sagar S. Patel, Taryn Cuper,

Robert K. Armstrong and Robert J. Alpino

School of Health Professions, Eastern Virginia Medical School, Norfolk, VA 23507, U.S.A.

Keywords: Digital Patient, Systems Biology, Physics-Based Modeling, Artificial Intelligence, Bigdata, Bio-Informatics,

Pharmaceutical Research, Virtual Clinical Trials, Personalized Medicine.

Abstract: The Digital Patient is an analytic platform that has the potential to transform personal and public healthcare,

pharmaceutical research, medical device development, and patient and professional education. It is the

ultimate big data project in healthcare; however, its power will derive not from the volume of data, but from

the successful and efficient integration of disparate sources of data into a validated and reliable computational

model of combined biological processes, social context and treatment efficacy. That integration, successively,

is largely dependent on the evolving theoretical approaches known as systems biology and physics-based

modeling that lead to the successful meshing of multi-scale models.

1 WHAT IS A DIGITAL PATIENT?

The Digital Patient is a comprehensive approach to

providing a computational platform for personalized

medicine. It is a digital representation of a person’s

‘health’ and ‘disease’ status or in another word, a

whole-body system. It may include computer models

of the mechanical, physical, and biochemical

functions of a living human body calibrated by multi-

scale and length data collected from the multiple

physiological levels. This is not only the ultimate big

data project, but also the ultimate technical challenge

in medicine. When executed, it can be used as a

powerful in silico (in silico means carried out in the

computer, which is in contrast to in vitro (on the

bench), ex vivo (outside the living organism), or in

vivo (inside the living organism)) decision support

technology that can be customized to represent each

one of us, individually and/or collectively (C. D.

Combs et al., 2015; Dıaz-Zuccarini et al., 2015;

Parodi, 2015). It is an integrated approach for

achieving a broader, more systematic understanding

of the human body in a single computational platform.

One of the key aspects to the project is the need

for a new analytic framework for understanding the

whole-body, that is, being able to deal with the

complexity of physiology by creating and integrating

Corresponding author

properly annotated, curated, validated and

documented modules of individual organ systems.

Ultimately, more complex models are expected to

merge upon linking these modules semantically to

represent the whole-body system. In the construction

of a Digital Patient framework, significantly

important things to consider are what level of detail is

necessary, and more importantly, how to accurately

predict the efficacy of a patient-specific treatment. In

addition, the processes for achieving a Digital Patient

framework involve generation of data and

information, biomedical information management,

knowledge-driven and data-driven computational

modeling (i.e., in silico models based on prior

knowledge on cause-and-effect relationships, and/or

nature of data), clinical user interface, and end

translation and adoption.

1.1 Integrative Whole-Body View

The human body is the ultimate efficient machine

comprised of multiple organs, tissues, and cellular

complexes that interact to maintain a homeostasis.

Pathology is due to alterations in one or more tissues

or systems or in a single biological process (Talbot et

al., 2016). Alterations in one system often induce

changes in the physiology of other systems.

Combs, C., Pinky, L., Ranathunge, C., Patel, S., Cuper, T., Armstrong, R. and Alpino, R.

Developing a Framework for Multi-Scale Modeling of the Digital Patient: Insights from Current Status and Future Directions.

DOI: 10.5220/0012140700003546

In Proceedings of the 13th International Conference on Simulation and Modeling Methodologies, Technologies and Applications (SIMULTECH 2023), pages 143-158

ISBN: 978-989-758-668-2; ISSN: 2184-2841

 2023 by SCITEPRESS – Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)

143

Therefore, the development of an integrated model of

human physiology is essential for the understanding

of how molecular, cellular, organ and system levels

interact for a total physiological response (An &

Cockrell, 2022; Hester et al., 2011). However,

biologists have traditionally sought to understand

living entities by investigating their constituent parts

in a controlled environment, i.e., in a reductionist

way, popularly known as reductionism (Morchio,

1991; Schaffner, 1976; Woese, 2004). For example,

first they isolated the individual genes, proteins, or

signaling molecules, and then they studied them

individually to learn everything they could about the

structure and function of that single biological entity,

without necessarily considering how they interact

with one another (Bertolaso, 2022; Bricmont, 2022;

Brigandt, 2013; Kuijper, 2022; Mazzocchi, 2008). In

contrast, the Digital Patient concept is based on an

integrated approach for achieving a broader, and

more systematic understanding of the human body,

and how it interacts with its environment, i.e., social

and behavioral factors (e.g., age, gender, body

weight, genetic make-up, lifestyle, etc. (C. D. Combs

& Combs, 2014).

Future healthcare will rely more and more on data

from monitoring devices and in some cases,

implanted medical devices such as wearable

biosensors. Interpretation of data, and their

therapeutic application requires knowledge that is

much more integrated and personalized than is

currently available (C. D. Combs & Combs, 2014;

Hatzikirou et al., 2012; D. P. Nickerson et al., 2015;

D. P. Nickerson et al., 2020; Tolk et al., 2015). Most

chronic diseases involve multiple organ systems.

Therefore, it is crucial to understand how the body

works as an integrative whole during homeostasis,

and at the same time to be able to utilize our detailed

knowledge of individual organs. Moreover, in order

to represent a comprehensive Digital Patient, social

systems and environmental factors must ultimately be

integrated into this analytic framework (C. D. Combs,

2017).

1.2 Biology View

Focusing more directly on converging different levels

of biological systems that are essential to the Digital

Patient framework is the discipline of systems

biology and its applications. Systems biology is an

integrative and interdisciplinary approach in contrast

to the traditional reductionist nature of biology

(Bertolaso, 2022; Bricmont, 2022; Brigandt, 2013;

Kuijper, 2022; Mazzocchi, 2008; Schaffner, 1976;

Woese, 2004). It attempts to explain complex

biological systems that include biochemical systems

(e.g., enzyme activity regulation and flux in

metabolic pathways), cellular processes (e.g., gene

regulation, protein transport, signaling pathways, the

cell cycle, and apoptosis), cell-cell interaction (e.g.,

cell–cell signaling), as well as cell differentiation and

organismal development (Boogerd et al., 2007;

Kitano, 2001; Klipp et al., 2016), using a variety of

conceptual and experimental methods such as

genomics, transcriptomics, proteomics, molecular

biology, cell biology, and carefully developed animal

models. Thus, systems biology has the potential to

provide valuable insights into the physiological

workings of the human body. The current goal of

systems biology research is to utilize scientific

advancements from the past two decades, such as

genomics and proteomics, in an effort to develop

targeted therapeutic strategies (Fitzgerald et al., 2006;

Khoo et al., 2021; Kohl et al., 2010). Over the past

two decades, sequencing technologies (e.g., Next-

Generation Sequencing, Whole-Genome Sequencing,

and etc.) have made remarkable progress (Hartman et

al., 2019). As these new technologies continue to

develop, the costs associated with sequencing have

decreased dramatically, making these technologies

more affordable and accessible. Rapid advances in

high-throughput technologies coupled with the

decrease in sequencing costs have led to generation

of massive amounts of biological data, and in turn, the

abundance of biological data has made data

integration approaches increasingly popular in the

field of systems biology

Systems biology not only addresses interactions

in biological systems at different scales of biological

organization, but also is characterized by quantitative

descriptions of biological processes, using a variety

of statistical and computational techniques (Baccam

et al., 2006; Karr et al., 2022; Perelson et al., 1996).

As was stated before, biological systems consist of a

large number of functionally diverse components,

which interact highly selectively and often

nonlinearly to produce coherent behaviors (Klipp et

al., 2005; Likić et al., 2010). These components may

be individual molecules (e.g., signaling or metabolic

networks), assemblies of interacting complexes, sets

of physical factors that guide the development of an

organism (genes, mRNA, associated proteins and

protein complexes), cells in tissues or organs, and

even entire organisms in ecological communities.

What is common to all these examples is the sheer

number of components, and their selective, non-linear

interactions that render the behaviors of these systems

beyond the intuitive grasp. Mathematical models of

biological systems are most suitable and are

SIMULTECH 2023 - 13th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

144

increasingly being used to represent our knowledge

about these systems (Iglesias & Ingalls, 2010; Ingalls,

2013). Thus, systems biology combines the

development and application of predictive

mathematical and computational modeling with

experimental studies. The quantitative techniques,

such as high-resolution microscopy, mass

spectrometry, flow cytometry, and more, that are

employed to incorporate multiple spatial and

temporal scales are consistent with the integrative

perspective of the Digital Patient framework.

1.3 Physics-Based Multi-Scale

Modeling View

Multi-scale modeling (MSM) integrates multiple

physiological processes across different length and

time scales to provide improved predictive and

individualized healthcare. The highly complex nature

of biomedical systems resulting from several distinct

factors includes the non-linearity and redundancy of

physiological states. They arise from multiple

mechanisms simultaneously pushing and pulling on

clinically relevant and/or experimentally observable

response variables (Vieira & Laubenbacher, 2022).

The concept of non-linearity states that many high-

level and integrative behaviors of the biological

system cannot be described solely through the sum of

inputs from basic processes (Auslander et al., 1972;

Oster & Perelson, 1973). The resulting heterogeneity

along with the disparate time constraints further

stimulates individual variability leading to distinct

disease outcomes across the population. Despite the

extensive complexities of biological and biomedical

systems, researchers are using both linear and non-

linear sophisticated biological and physiological

models to better understand fundamental relationships

within the human body (Bauer et al., 2009;

Beauchemin & Handel, 2011; Hester et al., 2011).

MSM, also known as knowledge-driven

modeling, mechanistic modeling, hypothesis- based

modeling, or physics-based modeling, is an equation-

based approach based on ordinary differential

equations, partial differential equations, stochastic

processes, agent-based modeling, etc. that

incorporates nonlinear coupled processes that occur

at different temporal and/or spatial scales, and lead to

a systematic integration of knowledge at the

molecular, cellular, and tissue levels (Altan-Bonnet et

al., 2020; Coveney & Fowler, 2005; Perelson &

Weisbuch, 1997; Pinky et al., 2021; Pinky &

Dobrovolny, 2017). Since biological entities have a

complex hierarchy of structure, mechanical

properties, and behavior across spatial and temporal

scales, MSM supports this integrative view by

explicitly defining the biological hypothesis or the

primary mechanisms and formalizing it into

mathematical equations (Harline et al., 2021; Pruett

& Hester, 2015; Radhakrishnan, 2020). This

approach assumes that behavior at larger scales

emerges naturally from the processes occurring at

smaller scales; in other words, embedding processes

at a small scale into the larger scales leads to a

prediction of overall system behavior (Gold et al.,

2019; Meier-Schellersheim et al., 2009). The

relationship between the biological and mathematical

theory determines the balance of phenomenology and

quantitative prediction. This step may become

iterative as the modeler balances the complexity of

the biological inputs included in the system with the

level of mathematical theory most suitable to form a

“minimum model” Kamerlin & Warshel, 2011;

Laubenbacher et al., 2021; Radhakrishnan, 2021;

Sego et al., 2022).

Working through biological and mathematical

theory should also reveal what data can be collected

from the system to inform the model construction,

i.e., what mechanisms and relationships are known to

exist, and which will be inferred, and what outcomes

or predictions will be tested. In addition, they can take

into account the influence of behavioral and social

context on the whole biological system. Thus, the

model connects the association of genetics to

proteins, proteins to cells, cells to organs, organs to

complete whole-body systems, as well as systems to

the organism itself and to the surrounding social

environment. For example, social and behavioral

context refers to taking into account the

understanding of the impact of the behavior of family

and friends on individual lifestyle choices and health

(C. D. Combs et al., 2015; Tolk et al., 2015).

1.4 Artificial Intelligence View

Artificial Intelligence (AI) is ideally suited to

discovering meaningful patterns in big data that may

otherwise escape human attention, and can offer a

more efficient means of understanding systems

dynamics and hence structuring preventive care

strategies more efficiently. Machine learning (ML)

method, a subset of AI tools, is at its core a data-

driven process, and defined as in silico models that

develop a predictor automatically for the data without

making any causal assumptions (Alber et al., 2019;

Peng et al., 2021). Unlike ML, MSM is generally

considered to be a theory-driven process and a more

traditional approach. It starts with developing

hypotheses, followed by collecting and analyzing data

Developing a Framework for Multi-Scale Modeling of the Digital Patient: Insights from Current Status and Future Directions

145

to test these hypotheses and drawing theoretical

conclusions based on the results (Pruett & Hester,

2015). It focuses on identifying abstract constructs

and the relationships among them. However, due to

the complexity of the environments and processes that

generate data, there may not be a strong theoretical

basis for the questions being studied ( T e i c h e r t

et al.,2019). In contrast, data-driven research

involves analyzing data to extract scientifically

interesting insights (i.e., robust correlations between

sets of variables) by applying analytical techniques

and modes of reasoning based on the data available

rather than prediction based on theory. It is worth

noting that some instances of ML in the literature are

described as theory-guided and seek causality by

integrating physics-based mechanistic models at

multiple temporal and spatial scales with big data

(Giansanti, 2022). In this way, ML can make up for

any unknown physics by learning the dynamics of the

system overall and thereby possibly classify patients

into specific treatment regimens. However, this

approach benefits from the knowledge and

mechanistic insights achieved through MSM to

develop novel learning algorithms with greater

robustness, data-efficiency, and generalization of

performance in data-limited situations (Peng et al.,

2021). As such, perhaps this approach can be

described as a meeting of ML and MSM to optimize

the contributions of both techniques. There are many

examples of data-driven MSM that appear to involve

the use of ML to optimize the parameter estimation

and functions. Perhaps these cases can also be

described as precursors to the combination of MSM

and ML (Alber et al., 2019; Maass et al., 2018).

In general, MSM provides insight into biological,

biomedical, and behavioral systems at a high level of

resolution and precision, which naturally produces

massive output data sets. Because computational

physiological simulations, e.g., physics-based MSM,

are too slow for clinical application, AI tools can

provide ways to speed up Digital Patient workflows.

For example, using ML methods one can develop a

simplified surrogate model (i.e., a statistical model to

accurately approximate simulation output) to reduce

complexity (Kennedy & O’Hagan, 2001). In the

context of the Digital Patient, MSM and ML

complement one another with respect to biological,

biomedical, and behavioral research and are possibly

even more powerful when combined (Costello &

Martin, 2018; Linka et al., 2020; Muzio et al., 2021).

The most sophisticated Digital Patient models are

expected to be self-improving. These models can

continuously monitor divergence between predictions

and observations, and use these divergences to

improve their own accuracy. Their deployment would

enable rapid refinement and improvement, especially

if they were designed in a modular fashion to permit

the parallel development and optimization of their

component sub-models (Maass et al., 2018).

1.5 Personalized Medicine View

Physiology is a basic medical science as knowledge

of normal functions of the body is the basis for

understanding diseases and identifying targets for

effective treatment (Sherwood, 2015). Just as

physiology is a branch of biology, systems

physiology, systems medicine and personalized

medicine are subsets of systems biology. Successful

responses to such a grand challenge, like the Digital

Patient, require this cross-disciplinary integration (An

& Cockrell, 2022; Grieves, 2019; P. Hunter et al.,

2002; Niarakis et al., 2022). Systems physiology

focuses on the function of interacting parts of the

system at the cell, tissue, organ and organ-system

scales, and is tightly coupled with structural

anatomical information (Sherwood, 2015). Systems

medicine is a subset of systems physiology that

addresses applications to clinical problems. Examples

include the application of the systems physiology

framework to develop quantitative understanding of

disease processes, leading to drug discovery, and to

the design of diagnostic tools (Tyson et al., 2001). A

subset of systems medicine that relies on individual

patient data or the data from a specific group of

similar patients is the emerging domain of

personalized medicine.

Realizing this goal requires the ability to make

accurate predictions about how a patient will react to

a treatment or no treatment; however, the previous

reductionist approach to science and modeling cannot

satisfy that need. Instead, the systems approach to

biological modeling is growing in importance as a

translational tool in clinical practice (Doyle III et al.,

2014). The explosion of data over the past twenty

years is providing novel opportunities to develop new

clinical treatments. New technologies such as DNA

sequencing, imaging, and proteomics provide

massive amounts of new information about the

human body. Further, these data can now be analyzed

at bulk or at the single-cell level, which has been

particularly useful to assess the tissue-level

heterogeneity in many diseases including cancer. The

ability to extract useful information from these data

has begun to lead to custom treatments for diseases,

such as cancer (Goldenberg et al., 2019; Khoo et al.,

2021), infectious diseases (Castiglioni et al., 2021; de

Fátima Cobre et al., 2021), diabetes (Kavakiotis et al.,

SIMULTECH 2023 - 13th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

146

2017) and hematological and metabolic disorders (De

Bruyne et al., 2021). This will, in turn, help improve

the health of individuals by converting research

findings into diagnostic tools and procedures.

2 CHALLENGES IN

DEVELOPING A DIGITAL

PATIENT FRAMEWORK

A large-scale implementation of human health into a

digital format requires the construction and execution

of highly complex computer models composed of

several component submodels which span multiple

spatial and temporal scales (P. Hunter, 2020; Pan et

al., 2021). In addition, multi-scale data is needed to

build this computational representation of biological

processes of the whole-body under both disease-free

and diseased states. Thus, several pieces must be in

place to realize this interplay in a single

computational model. These pieces include

reductionist modeling at a variety of spatial and

temporal scales, the development of an ontology

allowing models to communicate with one another,

and finally, the creation of a top-level model that

allows reductionist models to be plugged in, creating

an integrated model framework for the testing and

generation of hypotheses. Different groups have

developed distinct philosophies for approaching these

challenges, but none has solved the problem

completely (Hussan et al., 2022). Several challenges

to building a Digital Patient framework have been

identified and are discussed below.

2.1 Lack of a High Throughput

Approach to Modeling

A Digital Patient describing the disease state and

treatment requires the development, validation, and

integration of numerous component submodels in the

context of a rapidly developing scientific

understanding of biological behaviors and continual

generation of new experimental and clinical data

(Laubenbacher et al., 2021; Masison et al., 2021).

Although individual laboratories around the world

may construct submodels, the development of a

comprehensive Digital Patient framework will

require modularity to ensure validation and

interoperability with one another. In this way, the

submodels will handle complexity with modules that

are properly annotated, curated, and documented, and

then linked semantically to establish more complex

models. Enabling such parallel development requires

a flexible simulation architecture that uses a multi-

scale map of all the relevant components of a patient’s

response to the disease, as well as responses to

available treatments s (An & Cockrell, 2022; Grande

Gutiérrez et al., 2021; Masison et al., 2021).

2.2 Lack of Model Reproducibility and

Transparency

It is essential that component models utilized in the

Digital Patient are reproducible and reusable.

Published models relevant to this discussion

demonstrate a lack of transparency in model

implementation (Baker, 2016; Fitzpatrick, 2019). Not

all published models are reproducible and hence not

reusable. Furthermore, a significant number of

published mathematical models are not experimentally

validated, making model extensions more difficult

(Blinov et al., 2021; Niarakis., et al., 2022).

2.3 Generation of Heterogeneous

High-Dimensional Data

Development of a multi-scale model of an organ

requires the collection of synchronous measurements

at multiple physiological scales. This includes omics

data from tissues and single cells, from diverse

experimental systems, including two-dimensional

(2D) and 3D cell cultures, in vivo and ex vivo animal

models, patients, and biophysical and structural data

from tissues and organs, combined with data

characterizing transport throughout the body.

Technologically this is very challenging if not

impossible (Vieira & Laubenbacher, 2022).

2.4 Lack of Standardization in Data

Collection and Model Specification

Ever growing uncoordinated and heterogeneous

formats of data that capture the various determinants

of our health from genomic sequences to behavioral

influence lack (Vieira & Laubenbacher, 2022).

Thus, the use of incompatible data structures

along with almost no standard model specification

and different software environments, make

distributed collaboration difficult (Lubbock & Lopez,

2021; Malik-Sheriff et al., 2020).

2.5 Lack of Effective Communication

and Collaboration Among

Biomedical Researchers

Building a useful Digital Patient requires improved

communication between clinicians, experimentalists,

Developing a Framework for Multi-Scale Modeling of the Digital Patient: Insights from Current Status and Future Directions

147

and modelers in order to create sufficiently credible

interchangeable computational models and/or tools

that have value in the clinic, such as mobile apps,

dedicated web pages and medical devices. Inevitably,

this has led to the inability to efficiently translate

basic science knowledge obtained from pre-clinical

studies into effective therapies (An & Cockrell, 2022;

Grieves, 2019. To date, relatively few clinical and

biological insights are currently translated into

computational models that could serve as building

blocks for the Digital Patient framework.

2.6 Health Information Management

Having clinical information in electronic form that is

computable has been a grand challenge for

biomedical informatics (Acosta et al., 2022).

Unfortunately, most health information still sits in

silos today and health information exchange for the

purpose of supporting care between organizations and

levels of care (e.g., hospital to primary care), has,

until very recently, been the exception rather than the

norm. It is fair to say that, to a large extent,

management of health information has encountered

the most variability when we consider other related

domains like bioinformatics, pharmaceutical research

and development and medical device technology in

the quest for integrated biomedicine. Post-hoc data

collection has been shown to be very expensive and

error-prone because data sources can be very diverse

and range from operational electronic health record

systems to well-structured longitudinal disease

registries and bio-banks. Therefore, capturing

structured and computable clinical data as part of

routine clinical practice is ideal as it may otherwise

impossible to capture the clinical context in which the

data were collected initially. In addition, effectively

managing the enormous amount of personalized data

requires the development of broadly accepted policies

addressing security, quality control, data mining, and

privacy protection. This represents another

fundamental challenge to completing the Digital

Patient.

2.7 Patient Data Management

Further complicating the construction of the Digital

Patient framework is the current lack of agreement on

how we categorize patient information. An individual

patient’s ideal data set includes molecular data,

clinical data, and social context data, and these data

sets are not often integrated in a manner that is

understandable or easily usable by patients or

healthcare providers or even by research domain

experts (Kondylakis et al., 2015). Developing a

consistent terminology and aggregation methodology

for this disparate data is therefore an additional

fundamental challenge to building the Digital Patient

framework. Privacy, synchronicity (the timeliness

with which models produce actionable information),

and clarity of data organization and analysis are also

fundamental challenges that must be addressed in

completing the Digital Patient platform. Recent

3 DEVELOPMENT IN BUILDING

THE DIGITAL PATIENT

FRAMEWORK

There are many collaborative and individual efforts

underway that address some of the issues important

to building out the Digital Patient framework.

Following is a summary of several past and current

efforts, and tools developed. This list is a selection

and not exhaustive.

Projects that contributed most significant progress

are the Physiome and the 12 Labours projects; briefly

described below, the Physiome Project is an umbrella

term that refers to human modelling with methods

accommodating cross-disciplinary science

(chemistry, biology, physics) and a breadth of

dimensional and temporal scale (sub-cellular to

organs, sub-microsecond to tens-of-years) (P. Hunter

et al., 2002). The International Union of Physiological

Sciences (IUPS) Physiome Project, established in

1993, focused on providing a “computational

framework for understanding human and other

eukaryotic physiology”. This effort resulted in

databases, markup languages, software for

computational models of cell functions, as well as

software for interacting with organ models, as was

described in P. J. Hunter & Borg, 2003; P. Hunter,

2004. The primary limitation with the Physiome

Project has been the lack of integration of the multiple

narrow-focused models that could, if successfully

integrated, lead to a comprehensive and integrative

model of human physiology (D. P. Nickerson et al.,

2015; Viceconti et al., 2008). Today this project has

been extended into the 12 Labours research effort

currently underway within the Auckland

Bioengineering Institute at the University of

Auckland in New Zealand.

3.1 12 Labours Project

Initiated by the Auckland Bioengineering Institute

(ABI), University of Auckland in New Zealand, the

SIMULTECH 2023 - 13th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

148

intent of 12 Labours project is to extend the Physiome

Project to clinic and home-based healthcare

applications (P. Hunter, 2020). The focus and goal of

12 Labours is to create the infrastructure to integrate

multi-scale models into a whole-body computational

physiology system. This will include a platform for

precision medicine and sensor-based health

monitoring. Migrating the efforts from a research

focus to a clinical focus will necessarily require

changes in design and execution, e.g., model

reduction strategies must be utilized in order to

increase the speed at which models can be analyzed

and output a useful result. Additional foci include

coupling the physiome to body sensors for real-time

data exchange, an energy based mathematical

framework for understanding physiology, and a new

semantic approach to physics based multi-scale

modeling. In addition, with the goal of integrating

medical data with predictive physics modeling, this

project includes the development of workflow

management systems, the identification of

technologies for clinical translation of those

workflows, and the infrastructure for deploying those

workflows in a clinical environment. Some of the

suitable tools for workflow and data management

initially include Snake Make, NextFlow, iRods,

Pennsieve (Rajagopal et al., 2022).

Through the NIH Stimulation Peripheral Activity

to Relieve Conditions (SPARC) program, ABI has

been developing scaffolds of high-level descriptions

of organ anatomy on a 3D coordinate system

framework to relate multi-species organ models

(Osanlouy et al., 2021). These frameworks provide a

single common reference frame, upon which one can

mouse, pig. In other words, the framework is

consistent across multiple species, and the scaffolding

method facilitates cross-species comparisons as well

as the analysis of variation within a population.

Likewise, sub-scaffolds are used to define individual

characteristics within these scaffolds. The concept of

the whole-body scaffold reflects the use of this logic

for personalized models for virtual clinical trials via a

workflow in which organs and systems could be

assembled into whole body reference coordinates.

3.2 Computational Frameworks

Modularity is a particularly important consideration

when model dependencies are involved as the

introduction and absence of models can throw off the

entirety of a simulation. Similarly, multiple scales

require a framework with internal compensation to

account for varying definitions of, e.g., time.

Following are two examples of recent frameworks

that have been developed specifically to address these

concerns and that may serve to better light the way

forward for a Digital Patient framework. The first is

focused on medical digital twins and is a multi-

institutional effort of the Universities of Connecticut,

Michigan, and Florida. The second supports global

resource management and is out of Johns Hopkins

APL. Despite unrelated disciplines, the designs are

strikingly similar. If this review proves useful, there

may be others that can be reviewed in a similar

fashion. Following are two example computational

platforms for the integration of multiple, disparate

models.

• A Modular Computational Framework for

Medical Digital Twins With a focus on digital

twins in a medical context, researchers from

University of Florida Health sought to

optimize modularity by eliminating the

potential pitfalls of model dependencies with

an open source, “digital twin” architecture

(Masison et al., 2021). Characterized as “hub

and spoke” and hereafter referred to as MCF

for Modular Computational Framework, it

includes four components: a runtime

configuration file, a global model state,

modules, and a simulation framework that

controls simulation runtime and provides data

structures and algorithms useful for the

development of modules. Modules are

individual models, which can be added and

taken away at will; each module must provide

a subclass that defines the data relevant to that

module to be stored in the model state, i.e., a

pure data API. As such, the model state houses

all potential model inputs and outputs,

providing an indirect connection between

modules. One model’s output is stored in the

model state and can only from there be

accessed as input by another model. This also

ensures that all model processes stay

contained within a particular module, thereby

providing a clear separation between the

model and the data. In one paper (Masison et

al., 2021), MCF was applied to an existing

dynamic computational model of the immune

response to a respiratory fungal pathogen to

illustrate the potential of extending it to a full

digital twin use case. The MCF simulator is

open-source and available at (Masison, Joseph

and Beezley, Jonathan and Mei, Yu and

Ribeiro, Henrique Assis Lopes and Knapp,

Adam C and Sordo Vieira, L and Adhikari,

Bandita and Scindia, Yogesh and Grauer,

Developing a Framework for Multi-Scale Modeling of the Digital Patient: Insights from Current Status and Future Directions

149

Michael and Helba, Brian and others, 2021).

Additional platform access information is

available in the reference.

• System Integration with Multiscale

Networks (SIMoN) SIMoN was built to

realize the complex inter-relationships that

exist between the different facets of global

resource management (Hughes et al., 2020).

Each of these facets is a domain unto itself,

and models use their own geospatial,

temporal, and

other scales. SIMoN is an open-

source framework built to allow these

disparate models to be coupled. In the

referenced paper, they include climate,

population, and food-energy-water systems.

Specifically, SIMoN is used to “integrate

models and data from disparate domains by

predicting water availability in 2050, as it

depends on population growth, climate

change, and corresponding increases in

demand for thermoelectric cooling". While

SIMoN is not a biomedical use-case, the

challenges of integrating multiple distinct

models with different processes and scales

remain the same, requiring a framework that is

modular and extendable. Interestingly, SIMoN

employs a similar approach to the previous

example, with a “broker” construct taking the

place of the model state. Each model exists

inside of wrapper to standardize its interface

with the broker. The broker performs the

transformations necessary to reconcile the

varying geo-spatial scales of each model so

that those models can be integrated

seamlessly. In addition, it handles all data

inputs and outputs between models; like MCF,

this broker prevents direct dependencies of

one model on another. The only requirement

is that the models agree on the “scope” of a

given scale, e.g., the entirety of the contiguous

United States is the geo-spatial scope. The

scope can be subdivided by an individual

model as needed with the assumption that the

sum of all subdivisions exactly equals the

scope, with no overlap. It is not hard to

imagine how the same principle might be

applied to temporal and possibly other scales.

SIMoN is available at (Hughes, Marisa and

Kelbaugh, Michael and Campbell, Victoria

and Reilly, Elizabeth and Agarwala, Susama

and Wilt, Miller and Badger, Andrew and

Fuller, Evan and Ponzo, Dillon and Arevalo,

Ximena Calderon and others, 2020).

3.3 Collaborative Organizational

Efforts

The challenges of maximizing value from big data are

being addressed by the U.S. National Institutes of

Health’s (NIH) BD2K program, through the

European Union’s Horizon 2020 initiative, through

the European Big Data Value Association and

through various Chinese Ministry of Science and

Technology initiatives. In addition, the challenge of

encouraging consistency in terminology, ontology,

and registries is being addressed through the

International Health Terminology Standards

Development Organization, the Simulation Industry

Standards Organization and the NIH Data Discovery

Index Consortium. Moreover, model construction

and interoperability are the foci of the Physiome and

12 Labours Project, and of the researchers that are

involved in several organizations, including the

Virtual Physiological Human Institute (VPHi), the

U.S. Interagency Modeling and Analysis Group, and

its companion group, the Multi-Scale Modeling

Consortium. More such organizational efforts are

listed below.

• Insigneo, based on the Institute for In Silico

Medicine, is a collaboration between the

University of Sheffield, UK and the Sheffield

Teaching Hospitals with a focus on clinical

translation of in silico medicine. The project

implements the ambition behind the European

VPHi program. This is the largest organization

in Europe dedicated to the development,

validation, and use of in silico medical

technology.

• InSilco, dedicated to coronary artery disease,

is an international multidisciplinary

consortium focused on in silico trials for drug

eluting bioresorbable vascular scaffold design,

development, and evaluation.

• Physiome Journal is dedicated to the

reproducibility and reusability of models.

Each article is linked to a primary publication

in a peer-reviewed journal and includes access

to the presented model itself. This effort is

supported by the IUPS, VPHi, the University

of Auckland, Digital Science, and more.

• IMAG stands for Interagency Modeling and

Analysis Group (IMAG) based on NIH, USA,

holds multi-scale Modeling (MSM)

Consortium. Their goal includes supporting

research funding for modeling and analysis of

biomedical, biological, and behavioral

systems.

• InSilicoWorld is a worldwide community of

SIMULTECH 2023 - 13th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

150

practice working towards wider adoption of in

silico trials in the biomedical industry that is

currently led by Marco Viceconti (Viceconti et

al., 2008).

• SimBios, lead by NIH and Stanford

University, is a center for physics-based

simulation of biological structures. It provides

SimTK, developed OpenSim, and publishes

Biomedical Computation Review. It has also

established an inventory of bio-sim tools

called SimBiome in 2017.

Tools that were developed by the organizational

efforts are listed below.

• Markup Language Standards have been

developed to adequately describe physical and

physiological properties and processes

including CellML (Lloyd et al., 2004, FieldML

(Chang et al., 2007), TissueML (J.Q. et al.,

2004), AnatML (J.Q. et al., 2004), PhysioML

(J.Q. et al., 2004), SBML (Hucka et al., 2003).

• DICOM: Digital Imaging and

Communications in Medicine (DICOM) is the

international standard for medical images and

related information (Lim & Zein, 2006). It

defines formats for medical images that can be

exchanged with the data and quality necessary

for clinical use.

• OpenSim is an open-source simulation

software that was developed by the Stanford

National Center for Simulation in

Rehabilitation Research (Delp et al., 2007).

• BioModels is a repository of mathematical

models representing biological systems and is

written in SBML (Li et al., 2010). Models

include signaling, protein-drug interactions,

metabolic pathways, epidemic, and more.

BioModels was developed by the Molecular

Networks team EMBL-EBI based in UK and

the SBML Team from Caltech, USA.

• SimTK is a free biomedical project hosting

platform for the biomedical computation

community (Project-hosting platform for the

biomedical computation community, n.d.). It

provides an easy data sharing, shared resource

tracking and an infrastructure for community

connection and growth.

• OpenEHR is an open standard specification

in health informatics that allows semantic

mapping annotations into EHR data storage

formats (openEHR, 2017). From 2010-2020,

OpenEHR has been deployed in Australia,

Brazil, Switzerland, Germany, Finland, UK,

Italy, Malta, Netherlands, Norway,

Philippines, Russia, Sweden, and Slovenia.

• BioUML is web-based integrated

environment for systems biology and

collaborative analysis of biomedical data

(Russian Science Foundation, 2002). It was

funded and initiated by the Russian Science

Foundation in 2002. The initial goal of

BioUML was common purpose visual

language for formal descriptions of the

structure and function of biological systems.

The long-range plan is to be a computational

platform for the VPH and digital patient.

BioUML spans a comprehensive range of

capabilities, including access to biological

databases, powerful tools for systems biology

(visual modelling, simulation, parameters

fitting and analyses), a genome browser,

scripting (R, JavaScript), and a workflow

engine. The architecture is plugin-based.

Users create a visual representation of a model

and BioUML automatically generates the code

to simulate the model behavior. The current

version generates code in Java and uses its

own simulation engines. To support

collaborative work, there is a central

authentication and authorization system.

BioUML is open-source, in continued

development, and is actively used.

• MLBox was developed by the collaboration

with the University of Miami’s Miller School

of Medicine, the Media and Information Lab

(MIL), Amazon Web Services, and the

OpeHealth Network to create MLBox.

MLBox is an automated machine learning

Python library to support the development of

digital twins that can take the place of patients

to better test treatments options (Das &

Cakmak, 2018). Inputs include data from

wearable sensors and other smart devices,

including biological, clinical, behavioral, and

environmental. These are collected over a

period of seven days and combined into a

“biological health algorithm”, which in turn

acts as a digital twin in treatment tests. The

MLBox platform is in Python and is device

agnostic, i.e., modular, which will allow input

types to adjusted based on individual needs

and constraints, as well as expanded in the

future as technology evolves. The initial focus

area is sleep apnea and its link to dementia and

heart disease and inputs include sleep patterns,

weight, environmental pressures, and stress

levels.

• promor was developed by researchers at the

Eastern Virginia Medical School. Promor is an

Developing a Framework for Multi-Scale Modeling of the Digital Patient: Insights from Current Status and Future Directions

151

open-source R package that streamlines

biomarker discovery from proteomics data

and builds predictive models of disease

diagnosis and/or prognosis with top protein

biomarker candidates (Ranathunge et al.,

2023).

3.4 Other Collaborative Efforts

Biomedical research groups worldwide are

employing “digital twin” technologies to realize the

promise of personalized medicine. For example,

digital twins of the human heart can improve

diagnosis, prognosis, and therapies (Martinez-

Velazquez et al., 2019). Developers expect that

automated workflows for generating cardiac digital

twins could serve as a blueprint for the generation of

other types of medical digital twins (Corral-Acero et

al., 2020). Although medical digital twins are much

more difficult to develop than those for engineered

devices, they have begun to find applications in

improving human health. Examples include the

“artificial pancreas” for type 1 diabetes patients

(Breton et al., 2020; Brown et al., 2019; Kovatchev,

2019). In the artificial pancreas model, a template

mathematical model of human glucose metabolism

and a closed-loop control algorithm modeling insulin

delivery and data from an implanted glucose sensor

are customized into a patient-specific digital twin that

continuously calculates insulin needs and drives an

implanted pump that adjusts blood insulin

concentrations. Additionally, pediatric cardiac digital

twins combine template models of the heart with

patient-derived clinical measurements to optimize

some heart surgeries (Shang et al., 2019) and assess

the risk of thrombosis (Grande Gutiérrez et al., 2021;

Kondratova et al., 2019). The ARCHIMEDES

diabetes model expands these technologies by

including models not only of the progression of

diabetes within individual patients but also of medical

diagnosis, treatments, and the functioning of the health

care system that is providing the treatment (Du et al.,

2013; Eddy & Schlessinger, 2003a, 2003b).

More recently, the NIH Maximizing Investigators

Research Award (MIRA) was awarded to Dr. Tomas

Helikar, Professor of Biochemistry, University of

Nebraska, Lincoln, for the further development of a

virtual immune system. The virtual immune system is

meant to increase the understanding of immune

related diseases as well as to speed up drug

development and the time-to-market timeline. The

first MIRA award resulted in the successful modeling

of CD4+ T cells, which stimulate other cells to fight

pathogens. This model encompasses four

mathematical approaches, three spatial scales, and

multiple tissues involved in immune response. The

project established a method for computationally

connecting multiple scales of the immune system

(Wertheim et al., 2021. The goal of this second MIRA

award is to expand the model to include more types

of cells, molecules, genes, and organs.

A large part of the focus will be on computational

cost-effectiveness, improving the speed and

efficiency of the model’s algorithms.

In Europe, Neurotwin was initiated and funded by

the EU Horizon 2020 on January 1, 2021. It is

currently led by Neuroelectrics in Spain. It seeks to

predict the effects of non-invasive stimulation for

treatment of neurological disorders, e.g., Alzheimer’s

disease and epilepsy. Two proof-of-concept clinical

trials are planned for 2022 and 2023 in order to refine

the application of this stimulation technology for the

conditions of Alzheimer’s disease and epilepsy. If

successful, the condition use cases will be extended

to multiple sclerosis stroke rehabilitation, depression,

and the effects of psychedelics in the future.

Neurotwin combines 30 minutes of MRI and 10

minutes of EEG to create a personalized digital twin

that captures a brain’s electrical activity and

simulations the brain’s main parts, including the

scalp, skull, cerebrospinal fluid, and gray and white

matter. The digital twin also includes neural mass

models, or computational models of the average

behavior neurons using a map of neural connections

(connectome). This digital twin will be used to

optimize the stimulation position or locations and

strength of current via a headcap.

These examples illustrate how current digital

twins can operate in real time to maintain health

continuously, or they can be used off-line to design

personalized medical interventions.

3.5 Community Efforts

Community efforts such as the Systems Medicine

Disease Map Project, COVID-19 Disease Map

Project (Mazein et al., 2018; Ostaszewski et al., 2019;

Ostaszewski et al., 2020) and Computational

Modeling in Biology Network (COMBINE) are

working to build such infrastructure, although much

work needs to be done to adapt those for use in Digital

Patient framework. To this end, these groups have

built a large-scale data repository (Kondylakis et al.,

2017; Kondylakis et al., 2018; Kondylakis et al., 2015;

Kouroubali et al., 2019).

SIMULTECH 2023 - 13th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

152

4 CONCLUSIONS AND FUTURE

DIRECTIONS

Data is everywhere now, being aggregated, analyzed,

and repackaged. We are in an era of Big Data, living

with the recognition that almost everything we do is

being captured as one or another type of data, with the

hope that all that data can be used to help us become

smarter, healthier, safer, and richer. We also

recognize that our privacy may be invaded and that

our risk for harm is increasing. It is in this broader

context that this article addresses one of the more

hopeful Big Data undertakings - that is, the

construction and deployment of the Digital Patient.

The capacity to measure one’s personal physiological

and social metrics, compare those metrics with the

metrics of millions of other humans, personalize

therapeutic interventions and measure the resulting

changes will ultimately realize the vision of

personalized medicine - wherein patients and their

providers will be able to detect disease at an earlier

age and provide optimal therapy based on the

characteristics of each individual and reduce adverse

responses to therapy. Similarly, pharmaceutical

companies will improve the process of drug discovery

and clinical trials. In this way, the healthcare

industry’s emphasis truly shifts from reaction to

disease to prevention of disease and promotion of

wellness. Implicit in this vision is the integration of a

sustained focus on improving the outcome measures

of healthcare-safety, effectiveness, patient-

centeredness timeliness, efficiency, and equity in

clinical practice. Underlying this focus is, of course,

the development and integration of multi-scale

models based on the understandings emerging from

systems biology.

While the application of physics-based models to

the Digital Patient are exciting and varied, several

substantial challenges face the community. The two

most critical needs are connecting the top-down and

bottom-up model approaches. Modeling languages

have been established separately, and so the

community must spend valuable time and effort

replicating work already done by other groups. This

represents a gross inefficiency in the development

process, and hampers cooperation between groups. It

is our belief that organizations such as, the industry-

academia-regulatory agency consortia Avicenna

(European Commission, University of Sheffield and

the consortium, 2013) and the Medical Device

Innovation Consortium (Research collaborations in

regulatory science, 2011) will provide the force to

consolidate these efforts into a unified whole. The

more troubling challenge is that of rigorous model

validation. The assumptions that underlie the model

induce a standard for evaluating the model. In the

case of a larger target such as tissue, organ, or an

individual, validation becomes a more difficult

concept to define. Intense inter-subject variations

exist in humans. A person even demonstrates

different physiological characteristics at different

ages, so the existence of a data set that represents a

target for validation is often in question. Population

modeling may be the key; by generating many

individuals, a class of subjects similar to a given

patient might be selected over a collection of

observable variables. Consideration of differences in

that population may suggest other observations to

make in the patient, establishing an iterative process

for matching an individual to a reasonable model.

This challenge is not unique to biological models; it

exists across all nonlinear dynamic models, and no

systematic solution has been accepted (Barlas, 1996;

Coveney & Fowler, 2005). That said, the potential for

improving the modeling of individual patients and the

strata of patients that is represented by the Digital

Patient is clearly worth pursuing.

Author Contributions: All authors have read and

agreed to the published version of the manuscript.

Funding: This work was supported by the Hampton

Roads Biomedical Research Consortium.

Institutional Review Board Statement: Not

applicable. Informed Consent Statement: Not

applicable.

Data Availability Statement: Not applicable.

Conflicts of Interest: The authors declare no conflict

of interest. The funders had no role in the design of

the study; in the collection, analyses, or interpretation

of data; in the writing of the manuscript; or in the

decision to publish the results.

REFERENCES

Acosta, J. N., Falcone, G. J., Rajpurkar, P., & Topol, E. J.

(2022). Multimodal biomedical ai. Nature Medicine, 28

(9), 1773–1784.

Alber, M., Buganza Tepole, A., Cannon, W. R., De, S.,

Dura-Bernal, S., Garikipati, K., Karniadakis, G.,

Lytton, W. W., Perdikaris, P., Petzold, L., et al. (2019).

Integrating machine learning and multiscale

modeling—perspectives, challenges, and opportunities

in the biological, biomedical, and behavioral sciences.

NPJ Digital Medicine, 2 (1), 1–11.

Altan-Bonnet, G., Mora, T., & Walczak, A. M. (2020).

Quantitative immunology for physicists. Physics

Reports, 849, 1–83.

Developing a Framework for Multi-Scale Modeling of the Digital Patient: Insights from Current Status and Future Directions

153

An, G., & Cockrell, C. (2022). Drug development digital

twins for drug discovery, testing and repurposing: A

schema for requirements and development. Frontiers in

Systems Biology, 2.

Auslander, D., Perelson, A., Clifford, G., & Oster, G.

(1972). On systems with coupled chemical reaction and

diffusion. The American Society of Mechanical

Engineers.

Baccam, P., Beauchemin, C., Macken, C. A., Hayden, F.

G., & Perelson, A. S. (2006). Kinetics of influenza a

virus infection in humans. Journal of virology, 80 (15),

7590–7599.

Baker, M. (2016). Reproducibility crisis. Nature, 533 (26),

353–66.

Barlas, Y. (1996). Formal aspects of model validity and

validation in system dynamics. System Dynamics

Review: The Journal of the System Dynamics Society,

12 (3), 183–210.

Bauer, A. L., Beauchemin, C. A., & Perelson, A. S. (2009).

Agent-based modeling of host–pathogen systems: The

successes and challenges. Information Sciences, 179

(10), 1379–1389.

Beauchemin, C. A., & Handel, A. (2011). A review of

mathematical models of influenza a infections within a

host or cell culture: Lessons learned and challenges

ahead. BMC Public Health,11 (1), 1–15.

Bertolaso, M. (2022). Understanding complexity in life

sciences. In Environmental alteration leads to human

disease (pp. 1–13). Springer.

Blinov, M. L., Gennari, J. H., Karr, J. R., Moraru, I. I.,

Nickerson, D. P., & Sauro, H. M. (2021). Practical

resources for enhancing the reproducibility of

mechanistic modeling in systems biology. Current

Opinion in Systems Biology, 27, 100350.

Boogerd, F. C., Bruggeman, F. J., Hofmeyr, J.-H. S., &

Westerhoff, H. V. (2007). Towards philosophical

foundations of systems biology: Introduction. In

Systems biology (pp. 3–19). Elsevier.

Breton, M. D., Kanapka, L. G., Beck, R. W., Ekhlaspour,

L., Forlenza, G. P., Cengiz, E., Schoelwer, M., Ruedy,

K. J., Jost, E., Carria, L., et al. (2020). A randomized

trial of closed-loop control in children with type 1

diabetes. New England Journal of Medicine, 383 (9),

836–845.

Bricmont, J. (2022). Conclusion: Statistical mechanics and

reductionism. In Making sense of statistical mechanics

(pp. 311–314). Springer.

Brigandt, I. (2013). Systems biology and the integration of

mechanistic explanation and mathematical explanation.

Studies in History and Philosophy of Science Part C:

Studies in History and Philosophy of Biological and

Biomedical Sciences, 44 (4), 477–492.

Brown, S. A., Kovatchev, B. P., Raghinaru, D., Lum, J. W.,

Buckingham, B. A., Kudva, Y. C., Laffel, L. M., Levy,

C. J., Pinsker, J. E., Wadwa, R. P., et al. (2019). Six-

month randomized, multicenter trial of closed-loop

control in type 1 diabetes. New England Journal of

Medicine, 381 (18), 1707–1717.

Castiglioni, I., Ippolito, D., Interlenghi, M., Monti, C. B.,

Salvatore, C., Schiaffino, S., Polidori, A., Gandola, D.,

Messa, C., & Sardanelli, F. (2021). Machine learning

applied on chest x-ray can aid in the diagnosis of covid-

19: A first experience from lombardy, italy. European

Radiology Experimental, 5 (1), 1–10.

Chang, D., Lovell, N. H., & Dokos, S. (2007). Field markup

language: Biological field representation in xml. 2007

29th Annual International Conference of the IEEE

Engineering in Medicine and Biology Society, 402–

405.

Combs, C. D., Sokolowski, J. A., & Banks, C. M. (2015).

The digital patient: Advancing healthcare, research, and

education. John Wiley & Sons.

Combs, C. D. (2017). Developing a framework for multi-

scale modeling of the digital patient. International

Journal of Privacy and Health Information

Management (IJPHIM), 5 (2), 21–33.

Combs, C. D., & Combs, W. C. (2014). The digital patient:

An emerging platform for health analytics. Wiley

StatsRef: Statistics Reference Online, 1–9.

Corral-Acero, J., Margara, F., Marciniak, M., Rodero, C.,

Loncaric, F., Feng, Y., Gilbert, A., Fernandes, J. F.,

Bukhari, H. A., Wajdan, A., et al. (2020). The ‘digital

twin’ to enable the vision of precision cardiology.

European Heart Journal, 41 (48), 4556–4564.

Costello, Z., & Martin, H. G. (2018). A machine learning

approach to predict metabolic pathway dynamics from

time-series multiomics data. NPJ Systems Biology and

Applications, 4 (1), 1–14.

Coveney, P. V., & Fowler, P. W. (2005). Modelling

biological complexity: A physical scientist’s

perspective. Journal of the Royal Society Interface, 2

(4), 267–280.

Das, S., & Cakmak, U. M. (2018). Hands-on automated

machine learning: A beginner’s guide to building

automated machine learning systems using automl and

python. Packt Publishing Ltd.

De Bruyne, S., Speeckaert, M. M., Van Biesen, W., &

Delanghe, J. R. (2021). Recent evolutions of machine

learning applications in clinical laboratory medicine.

Critical Reviews in Clinical Laboratory Sciences, 58

(2), 131–152.

de Fátima Cobre, A., Stremel, D. P., Noleto, G. R., Fachi,

M. M., Surek, M., Wiens, A., Tonin, F. S., & Pontarolo,

R. (2021). Diagnosis and prediction of covid-19

severity: Can biochemical tests and machine learning

be used as prognostic indicators? Computers in Biology

and Medicine, 134, 104531.

Delp, S. L., Anderson, F. C., Arnold, A. S., Loan, P., Habib,

A., John, C. T., Guendelman, E., & Thelen, D. G.

(2007). Opensim: Open-source software to create and

analyze dynamic simulations of movement. IEEE

Transactions on Biomedical Engineering, 54 (11),

1940–1950.

Dıaz-Zuccarini, V., Alimohammadi, M., & Pichardo-

Almarza, C. (2015). Reflecting on discipulus and

remaining challenges. The Digital Patient: Advancing

Healthcare, Research, and Education, 15.

Doyle III, F. J., Huyett, L. M., Lee, J. B., Zisser, H. C., &

Dassau, E. (2014). Closed-loop artificial pancreas

SIMULTECH 2023 - 13th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

154

systems: Engineering the algorithms. Diabetes Care, 37

(5), 1191–1197.

Du, P., O’Grady, G., Gao, J., Sathar, S., & Cheng, L. K.

(2013). Toward the virtual stomach: Progress in

multiscale modeling of gastric electrophysiology and

motility. Wiley Interdisciplinary Reviews: Systems

Biology and Medicine, 5 (4), 481–493.

Eddy, D. M., & Schlessinger, L. (2003a). Archimedes: A

trial-validated model of diabetes. Diabetes Care, 26

(11), 3093–3101.

Eddy, D. M., & Schlessinger, L. (2003b). Validation of the

archimedes diabetes model. Diabetes Care, 26 (11),

3102–3110.

European Commission, University of Sheffield and the

consortium. (2013). Avicenna: A strategy for in silico

clinical trials [http://avicenna-isct.org/].

Fitzgerald, J. B., Schoeberl, B., Nielsen, U. B., & Sorger, P.

K. (2006). Systems biology and combination therapy in

the quest for clinical efficacy. Nature Chemical

Biology, 2 (9), 458–466.

Fitzpatrick, B. G. (2019). Issues in reproducible simulation

research. Bulletin of Mathematical Biology, 81 (1), 1–

Giansanti, D. (2022). Artificial intelligence in public health:

Current trends and future possibilities. Gold, K.,

Gaharwar, A. K., & Jain, A. (2019). Emerging trends in

multiscale modeling of vascular pathophysiology:

Organ-on-a-chip and 3d printing. Biomaterials, 196, 2–

17.

Goldenberg, S. L., Nir, G., & Salcudean, S. E. (2019). A

new era: Artificial intelligence and machine learning in

prostate cancer. Nature Reviews Urology, 16 (7), 391–

403.

Grande Gutiérrez, N., Alber, M., Kahn, A. M., Burns, J. C.,

Mathew, M., McCrindle, B. W., & Marsden, A. L.

(2021). Computational modeling of blood component

transport related to coronary artery thrombosis in

kawasaki disease. PLoS Computational Biology, 17 (9),

e1009331.

Grieves, M. W. (2019). Virtually intelligent product

systems: Digital and physical twins.

Harline, K., Martınez-G´omez, J., Specht, C. D., & Roeder,

A. H. (2021). A life cycle for modeling biology at

different scales. Frontiers in Plant Science, 1724.

Hartman, P., Beckman, K., Silverstein, K., Yohe, S.,

Schomaker, M., Henzler, C., Onsongo, G., Lam, H. C.,

Munro, S., Daniel, J., et al. (2019). Next generation

sequencing for clinical diagnostics: Five year

experience of an academic laboratory. Molecular

Genetics and Metabolism Reports, 19, 100464.

Hatzikirou, H., Chauviere, A., Bauer, A. L., Leier, A.,

Lewis, M. T., Macklin, P., Marquez-Lago, T. T.,

Bearer, E. L., & Cristini, V. (2012). Integrative physical

oncology. Wiley Interdisciplinary Reviews: Systems

Biology and Medicine, 4 (1), 1–14.

Hester, R. L., Iliescu, R., Summers, R., & Coleman, T. G.

(2011). Systems biology and integrative physiological

modelling. The Journal of Physiology, 589 (5), 1053–

1060.

Hucka, M., Finney, A., Sauro, H. M., Bolouri, H., Doyle, J.

C., Kitano, H., Arkin, A. P., Bornstein, B. J., Bray, D.,

Cornish-Bowden, A., et al. (2003). The systems biology

markup language (sbml): A medium for representation

and exchange of biochemical network models.

Bioinformatics, 19 (4), 524–531.

Hughes, M., Kelbaugh, M., Campbell, V., Reilly, E.,

Agarwala, S., Wilt, M., Badger, A., Fuller, E., Ponzo,

D., Arevalo, X. C., et al. (2020). System integration

with multiscale networks (simon): A modular

framework for resource management models. 2020

Winter Simulation Conference (WSC), 656–667.

Hughes, Marisa and Kelbaugh, Michael and Campbell,

Victoria and Reilly, Elizabeth and Agarwala, Susama

and Wilt, Miller and Badger, Andrew and Fuller, Evan

and Ponzo, Dillon and Arevalo, Ximena Calderon and

others. (2020). System integration with multiscale

networks [https://github.com/JHUAPL/SIMoN].

Hunter, P. (2020). Modeling framework for computational

physiology. Encyclopedia of Continuum Mechanics,

1691–1702.

Hunter, P., Robbins, P., & Noble, D. (2002). The iups

human physiome project. Pfl¨ugers Archiv,445 (1), 1–

Hunter, P. J., & Borg, T. K. (2003). Integration from

proteins to organs: The physiome project. Nature

Reviews Molecular Cell Biology, 4 (3), 237–243.

Hunter, P. (2004). The iups physiome project: A framework

for computational physiology. Progress in Biophysics

and Molecular Biology, 85 (2-3), 551–569.

Hussan, J. R., Trew, M. L., & Hunter, P. J. (2022).

Simplifying the process of going from cells to tissues

using statistical mechanics. Frontiers in physiology,

279.

Iglesias, P. A., & Ingalls, B. P. (2010). Control theory and

systems biology. MIT press.

Ingalls, B. P. (2013). Mathematical modeling in systems

biology: An introduction. MIT press.

J.Q., D., Q., L., & Q.M., L. (2004). Java and xml based

graphic software package for modeling biochemical

and gene regulatory networks. Computational

Engineering Application, 40 (6), 202–204.

Kamerlin, S. C. L., & Warshel, A. (2011). Multiscale

modeling of biological functions. Physical Chemistry

Chemical Physics, 13 (22), 10401–10411.

Karr, J., Malik-Sheriff, R. S., Osborne, J., Gonzalez-Parra,

G., Forgoston, E., Bowness, R., Liu, Y., Thompson, R.,

Garira, W., Barhak, J., et al. (2022). Model integration

in computational biology: The role of reproducibility,

credibility and utility. Frontiers in Systems Biology, 2.

Kavakiotis, I., Tsave, O., Salifoglou, A., Maglaveras, N.,

Vlahavas, I., & Chouvarda, I. (2017). Machine learning

and data mining methods in diabetes research.

Computational and Structural Biotechnology Journal,

15, 104–116.

Kennedy, M. C., & O’Hagan, A. (2001). Bayesian

calibration of computer models. Journal of the Royal

Statistical Society: Series B (Statistical Methodology),

63 (3), 425–464.

Developing a Framework for Multi-Scale Modeling of the Digital Patient: Insights from Current Status and Future Directions

155

Khoo, A., Liu, L. Y., Nyalwidhe, J. O., Semmes, O. J.,

Vesprini, D., Downes, M. R., Boutros, P. C., Liu, S. K.,

& Kislinger, T. (2021). Proteomic discovery of non-

invasive biomarkers of localized prostate cancer using

mass spectrometry. Nature Reviews Urology, 18 (12),

707–724.

Kitano, H. (2001). Foundations of systems biology. The

MIT Press Cambridge, Massachusetts London,

England.

Klipp, E., Herwig, R., Kowald, A., Wierling, C., &

Lehrach, H. (2005). Systems biology in practice:

Concepts, implementation and application. John Wiley

& Sons.

Klipp, E., Liebermeister, W., Wierling, C., & Kowald, A.

(2016). Systems biology: A textbook. John Wiley &

Sons.

Kohl, P., Crampin, E. J., Quinn, T., & Noble, D. (2010).

Systems biology: An approach. Clinical Pharmacology

& Therapeutics, 88 (1), 25–33.

Kondratova, M., Czerwinska, U., Sompairac, N.,

Amigorena, S. D., Soumelis, V., Barillot, E., Zinovyev,

A., & Kuperstein, I. (2019). A multiscale signalling

network map of innate immune response in cancer

reveals cell heterogeneity signatures. Nature

Communications, 10 (1), 1–13.

Kondylakis, H., Bucur, A., Dong, F., Renzi, C., Manfrinati,

A., Graf, N., Hoffman, S., Koumakis, L., Pravettoni, G.,

Marias, K., et al. (2017). Imanagecancer: Developing a

platform for empowering patients and strengthening

self-management in cancer diseases. 2017 IEEE 30th

International Symposium on Computer-Based Medical

Systems (CBMS), 755–760.

Kondylakis, H., Koumakis, L., Tsiknakis, M., & Marias, K.

(2018). Implementing a data management

infrastructure for big healthcare data. 2018 IEEE

EMBS International Conference on Biomedical &

Health Informatics (BHI), 361–364.

Kondylakis, H., Spanakis, E. G., Sfakianakis, S., Sakkalis,

V., Tsiknakis, M., Marias, K., Zhao, X., Yu, H. Q., &

Dong, F. (2015). Digital patient: Personalized and

translational data management through the

MyHealthAvatar EU project. 2015 37th Annual

International Conference of the IEEE Engineering in

Medicine and Biology Society (EMBC), 1397–1400.

Kouroubali, A., Koumakis, L., Kondylakis, H., &

Katehakis, D. G. (2019). An integrated approach

towards developing quality mobile health apps for

cancer. In Mobile health applications for quality

healthcare delivery (pp. 46–71). IGI Global.

Kovatchev, B. (2019). A century of diabetes technology:

Signals, models, and artificial pancreas control. Trends

in Endocrinology & Metabolism, 30 (7), 432–444.

Kuijper, H. (2022). Reductionism or holism? In

Comprehending the complexity of countries (pp. 247–

273). Springer.

Laubenbacher, R., Sluka, J. P., & Glazier, J. A. (2021).

Using digital twins in viral infection.Science, 371

(6534), 1105–1106.

Li, C., Donizelli, M., Rodriguez, N., Dharuri, H., Endler,

L., Chelliah, V., Li, L., He, E., Henry, A., Stefan, M. I.,

et al. (2010). Biomodels database: An enhanced,

curated and annotated resource for published

quantitative kinetic models. BMC Systems Biology, 4

(1), 1–14.

Likić, V. A., McConville, M. J., Lithgow, T., & Bacic, A.

(2010). Systems biology: The next frontier for

bioinformatics. Advances in Bioinformatics, 2010.

Lim, J., & Zein, R. (2006). The digital imaging and

communications in medicine (dicom): Description,

structure and applications. In Rapid prototyping (pp.

63–86). Springer.

Linka, K., Peirlinck, M., & Kuhl, E. (2020). The

reproduction number of covid-19 and its correlation

with public health interventions. Computational

Mechanics, 66 (4), 1035–1050.

Lloyd, C. M., Halstead, M. D., & Nielsen, P. F. (2004).

Cellml: Its future, present and past. Progress in

Biophysics and Molecular Biology, 85 (2-3), 433–450.

Lubbock, A. L., & Lopez, C. F. (2021). Programmatic

modeling for biological systems. Current Opinion in

Systems Biology, 27, 100343.

Maass, W., Parsons, J., Purao, S., Storey, V. C., & Woo, C.

(2018). Data-driven meets theory-driven research in the

era of big data: Opportunities and challenges for

information systems research. Journal of the

Association for Information Systems, 19 (12), 1.

Malik-Sheriff, R. S., Glont, M., Nguyen, T. V., Tiwari, K.,

Roberts, M. G., Xavier, A., Vu, M. T., Men, J., Maire,

M., Kananathan, S., et al. (2020). Biomodels—15 years

of sharing computational models in life science.

Nucleic Acids Research, 48 (D1), D407–D415.

Martinez-Velazquez, R., Gamez, R., & El Saddik, A.

(2019). Cardio twin: A digital twin of the human heart

running on the edge. 2019 IEEE International

Symposium on Medical Measurements and

Applications (MeMeA), 1–6.

Masison, J., Beezley, J., Mei, Y., Ribeiro, H. A. L., Knapp,

A. C., Sordo Vieira, L., Adhikari, B., Scindia, Y.,

Grauer, M., Helba, B., et al. (2021). A modular

computational framework for medical digital twins.

Proceedings of the National Academy of Sciences, 118

(20), e2024287118.

Masison, Joseph and Beezley, Jonathan and Mei, Yu and

Ribeiro, Henrique Assis Lopes and Knapp, Adam C and

Sordo Vieira, L and Adhikari, Bandita and Scindia,

Yogesh and Grauer, Michael and Helba, Brian and

others. (2021). A modular computational framework

for medical digital twins [https://github.com/

NutritionalLungImmunity/nlisim].

Mazein, A., Ostaszewski, M., Kuperstein, I., Watterson, S.,

Le Nov`ere, N., Lefaudeux, D., De Meulder, B., Pellet,

J., Balaur, I., Saqi, M., et al. (2018). Systems medicine

disease maps: Community-driven comprehensive

representation of disease mechanisms. NPJ Systems

Biology and Applications, 4 (1), 1–10.

Mazzocchi, F. (2008). Complexity in biology: Exceeding

the limits of reductionism and determinism using

complexity theory. EMBO Reports, 9 (1), 10–14.

Meier-Schellersheim, M., Fraser, I. D., & Klauschen, F.

(2009). Multiscale modeling for biologists. Wiley

SIMULTECH 2023 - 13th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

156

Interdisciplinary Reviews: Systems Biology and

Medicine, 1 (1), 4–14.

Morchio, R. (1991). Reductionism in biology. In The

problem of reductionism in science (pp. 149–160).

Springer.

Muzio, G., O’Bray, L., & Borgwardt, K. (2021). Biological

network analysis with deep learning. Briefings in

Bioinformatics, 22 (2), 1515–1530.

Narkhede, S. M., Luther, L., Raugh, I. M., Knippenberg, A.

R., Esfahlani, F. Z., Sayama, H., Cohen, A. S.,

Kirkpatrick, B., & Strauss, G. P. (2022). Machine

learning identifies digital phenotyping measures most

relevant to negative symptoms in psychotic disorders:

Implications for clinical trials. Schizophrenia Bulletin,

48 (2), 425–436.

Niarakis, A., Waltemath, D., Glazier, J., Schreiber, F.,

Keating, S. M., Nickerson, D., Chaouiya, C., Siegel, A.,

No¨el, V., Hermjakob, H., et al. (2022). Addressing

barriers in comprehensiveness, accessibility,

reusability, interoperability and reproducibility of

computational models in systems biology. Briefings in

Bioinformatics.

Nickerson, D. P., Atalag, K., de Bono, B., & Hunter, P. J.

(2015). The physiome project, openehr, archetypes, and

the digital patient. The Digital Patient: Advancing

Healthcare, Research, and Education, John Wiley &

Sons, 101–124.

Nickerson, D. P., Lundeng˚ard, K., Watts, J., Porter, S., Yu,

T., Nielsen, P., & Hunter, P. (2020). Physiome:

Encouraging reproducible and fair computational

modelling. The FASEB Journal, 34 (S1), 1–1.

openEHR. (2017). An open domain-driven platform for

developing flexible e-health systems.

Osanlouy, M., Bandrowski, A., De Bono, B., Brooks, D.,

Cassar`a, A. M., Christie, R., Ebrahimi, N., Gillespie,

T., Grethe, J. S., Guercio, L. A., et al. (2021). The sparc

drc: Building a resource for the autonomic nervous

system community. Frontiers in Physiology, 12.

Ostaszewski, M., Gebel, S., Kuperstein, I., Mazein, A.,

Zinovyev, A., Dogrusoz, U., Hasenauer, J., Fleming, R.

M., Le Novere, N., Gawron, P., et al. (2019).

Community-driven roadmap for integrated disease

maps. Briefings in Bioinformatics, 20 (2), 659–670.

Ostaszewski, M., Mazein, A., Gillespie, M. E., Kuperstein,

I., Niarakis, A., Hermjakob, H., Pico, A. R.,

Willighagen, E. L., Evelo, C. T., Hasenauer, J., et al.

(2020). COVID-19 disease map, building a

computational repository of SARS-CoV-2 virus-host

interaction mechanisms. Scientific Data, 7 (1), 1–4.

Oster, G. F., & Perelson, A. S. (1973). Systems, circuits and

thermodynamics. Israel Journal of Chemistry, 11 (2-3),

445–478.

Pan, M., Gawthrop, P. J., Cursons, J., & Crampin, E. J.

(2021). Modular assembly of dynamic models in

systems biology. PLoS Computational Biology, 17

(10), e1009513.

Parodi, V. A. (2015). The digital patient: Changing the

paradigm of healthcare and impacting medical research

and education. The Digital Patient: Advancing

Healthcare, Research, and Education, 273–88.

Peng, G. C., Alber, M., Buganza Tepole, A., Cannon, W.

R., De, S., Dura-Bernal, S., Garikipati, K., Karniadakis,

G., Lytton, W. W., Perdikaris, P., et al. (2021).

Multiscale modeling meets machine learning: What can

we learn? Archives of Computational Methods in

Engineering, 28 (3), 1017–1037.

Perelson, A. S., Neumann, A. U., Markowitz, M., Leonard,

J. M., & Ho, D. D. (1996). Hiv-1 dynamics in vivo:

Virion clearance rate, infected cell life-span, and viral

generation time. Science, 271 (5255), 1582–1586.

Perelson, A. S., & Weisbuch, G. (1997). Immunology for

physicists. Reviews of Modern Physics, 69 (4), 1219.

Pinky, L., Burke, C. W., Russell, C. J., & Smith, A. M.

(2021). Quantifying dose-, strain-, and tissue-specific

kinetics of parainfluenza virus infection. PLoS

Computational Biology, 17 (8), e1009299.

Pinky, L., & Dobrovolny, H. M. (2017). The impact of cell

regeneration on the dynamics of viral coinfection.

Chaos: An Interdisciplinary Journal of Nonlinear

Science, 27 (6), 063109.

Project-hosting platform for the biomedical computation

community. (n.d.). Web-based integrated environment

for systems biology [https://simtk.org/

whatIsSimtk.php].

Pruett, W. A., & Hester, R. L. (2015). Physics-based

modeling for the physiome. The Digital Patient:

Advancing Healthcare, Research, and Education, 127.

Radhakrishnan, R. (2020). Multiscale modeling:

Foundations, historical milestones, current status, and

future prospects. Authorea Preprints.

Radhakrishnan, R. (2021). A survey of multiscale

modeling: Foundations, historical milestones, current

status, and future prospects. AIChE Journal, 67 (3),

e17026.

Rajagopal, V., Arumugam, S., Hunter, P., Khadangi, A.,

Chung, J., & Pan, M. (2022). The cell physiome: What

do we need in a computational physiology framework

for predicting single cell biology? arXiv preprint

arXiv:2202.13282.

Ranathunge, C., Patel, S. S., Pinky, L., Correll, V. L., Chen,

S., Semmes, O. J., Armstrong, R. K., Combs, C. D., &

Nyalwidhe, J. O. (2023). promor: a comprehensive R

package for label-free proteomics data analysis and

predictive modeling [vbad025]. Bioinformatics

Advances,3 (1). https://doi.org/10.1093/bioadv/

vbad025

Research collaborations in regulatory science. (2011).

Medical device innovation consortium [http://mdic.

org/].

Russian Science Foundation. (2002). Web-based integrated

environment for systems biology [https://bio-store.org].

Schaffner, K. F. (1976). Reductionism in biology:

Prospects and problems. PSA 1974, 613–632.

Sego, T., Mochan, E. D., Ermentrout, G. B., & Glazier, J.

A. (2022). A multiscale multicellular spatiotemporal

model of local influenza infection and immune

response. Journal of Theoretical Biology, 532, 110918.

Shang, J. K., Esmaily, M., Verma, A., Reinhartz, O.,

Figliola, R. S., Hsia, T.-Y., Feinstein, J. A., & Marsden,

A. L. (2019). Patient-specific multiscale modeling of

Developing a Framework for Multi-Scale Modeling of the Digital Patient: Insights from Current Status and Future Directions

157

the assisted bidirectional glenn. The Annals of Thoracic

Surgery, 107 (4), 1232–1239.

Sherwood, L. (2015). Human physiology: From cells to

systems. Cengage learning. Smith, A. M. (2018).

Validated models of immune response to virus

infection. Current Opinion in Systems Biology, 12, 46–

52.

Talbot, S., Foster, S. L., & Woolf, C. J. (2016).

Neuroimmunity: Physiology and pathology. Annual

Review of Immunology, 34, 421–447.

Teichert, G. H., Natarajan, A. R., Van der Ven, A., &

Garikipati, K. (2019). Machine learning materials

physics: Integrable deep neural networks enable scale

bridging by learning free energy functions. Computer

Methods in Applied Mechanics and Engineering, 353,

201–216.

Tolk, A., Balci, O., Combs, C. D., Fujimoto, R., Macal, C.

M., Nelson, B. L., & Zimmerman, P. (2015). Do we

need a national research agenda for modeling and

simulation? 2015 Winter Simulation Conference

(WSC), 2571–2585.

Tyson, J. J., Chen, K., & Novak, B. (2001). Network

dynamics and cell physiology. Nature Reviews

Molecular Cell Biology, 2 (12), 908–916.

Viceconti, M., Clapworthy, G., & Jan, S. V. S. (2008). The

virtual physiological human—a european initiative for

in silico human modelling—. The Journal of

Physiological Sciences, 0810200082–0810200082.

Vieira, L. S., & Laubenbacher, R. C. (2022). Computational

models in systems biology: Standards, dissemination,

and best practices. Current Opinion in Biotechnology,

75, 102702.

Wertheim, K. Y., Puniya, B. L., La Fleur, A., Shah, A. R.,

Barberis, M., & Helikar, T. (2021). A multi-approach

and multi-scale platform to model cd4+ t cells

responding to infections. PLoS Computational Biology,

17 (8), e1009209.

Woese, C. R. (2004). A new biology for a new century.

Microbiology and Molecular Biology Reviews, 68 (2),

173–186.

SIMULTECH 2023 - 13th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

158