I hold an M.Sc. in theoretical physics and completed my PhD on the modeling of biological systems at ISY (Link枚ping University) and S2 (Chalmers). Subsequently, I established my own modeling group as a subgroup within an experimental cell biology group. The group expanded, and in 2011 we became an independent research group here at IMT; we now comprise approximately 20 staff members and ten students. Since then, we have continued to grow; I have been appointed visiting professor at 脰rebro University, and I now also lead several major national and international networks.
Our modeling approach: hypothesis testing and digital twins
As a standard procedure in our modeling projects, we take experimental data and biomedical knowledge from our collaborators and formulate mechanistic hypotheses. These hypotheses are translated into mathematical models (typically using ordinary differential equations) that are then fitted to training data. This leads to one of two outcomes:
- the hypothesis cannot explain the data and requires reformulation, or
- the hypothesis can describe the training data and is subsequently tested against independent validation data
allowing the model to be used, for instance, to design new experiments where the model's predictions can be further tested. In this way, we become part of daily decision-making in the experimental environment, helping to analyze data and plan new experiments.
We have employed this strategy to characterize various aspects of most major organs in the body: adipose and muscle tissue, liver, brain, pancreas, blood, etc. We are currently linking these organ models to enable whole-body modeling. These interconnected models can be individualized by fitting them to personal data. Such individualized models are sometimes referred to as digital twins which can serve a wide range of purposes. We have also developed similar models for other complex experimental systems, such as organs-on-a-chip and rodent models. These models span scales from intracellular processes to organs and whole organisms, and timeframes ranging from seconds to years. In other words, our interconnected models can be described as "M4" models: mechanistic, multi-level, multi-timescale, and multi-species.
Applications of digital twins in healthcare, prevention, and drug development
These digital twins, or M4 models, are useful in a variety of contexts. They can be employed to promote human health (鈥渕odels for health,鈥� or 鈥淢4-health鈥�), for instance, through the development of e-health products at our spin-off company, SUND Sound Medical Decisions. In the initial phase, we are testing whether these digital twins can enhance health education and boost patient motivation to adhere to treatment plans and adopt preventive measures. The digital twin accompanies the patient from the initial health consultation through to the use of smart sensors for monitoring and recording health status at home. Finally, it plays a role in specialist care for conditions such as type 2 diabetes or liver and heart complications.
Beyond these applications, we are collaborating with AstraZeneca to transform drug development from a linear trial-and-error pipeline into a knowledge-driven workflow. This approach holds great potential for replacing animal testing, an initiative we pursue in partnership with the organization 鈥淔orska Utan Djurf枚rs枚k鈥� (Research Without Animal Testing) and the Swedish 3R Center, which I helped to setup.
Lastly, we are now also utilizing our digital twin in innovative art projects, combining my piano playing with professional dancers and dancing digital twins in a novel format for lecture-performances.
Leading large national and international networks
The development of my digital twins began as individual projects and collaborations 鈥� supported, for instance, by research grants from the Swedish Research Council (VR) 鈥� and has since evolved into major network projects. One of the first such projects I coordinated was funded by VINNOVA and involved Region 脰sterg枚tland and Lund University.
This subsequently led to several EU projects: first PRECISE4Q (2019鈥�2022) and then STRATIF-AI (2023鈥�2027). I coordinate the latter, which comprises 15 partners from eight countries and has a budget of six million euros. This project focuses on how my digital twins can be used for stroke prevention, acute treatment, and rehabilitation, and involves testing e-health apps on over 400 individuals across six countries.
M4-HEALTH the largest and most exciting network
However, the largest and most exciting network I have initiated is M4-HEALTH. It consists of 16 Swedish universities, engages key international universities and networks, and is supported by over 100 companies, government agencies, and other organisations. We have submitted an application to become a cluster of excellence, but we are also pursuing other similar funding schemes and will continue working on this network in one way or another regardless of funding schema.
Through M4-HEALTH, we are adopting a holistic approach to health that has previously been lacking in society: we integrate all knowledge and data into the digital twin, and we bring together all stakeholders involved in health, not just healthcare and elderly care, but also the defense sector, schools, the food industry, and so on.
This holistic approach has the potential to make a real difference. We see the opportunity and the importance of reversing the trend towards unhealth, which has been going on for decades, and to instead embark on a new, vision-driven journey towards improved health and well-being in society. Enabling and leading this journey is the overarching goal of all of my research.
