An £11m digital twins centre will model organs and disease to speed drug discovery and help target medicines more precisely.
The Modelling-Informed Medicine Centre, or MiMeC, will build computer models of organs and diseases to better understand how diseases of the lungs, liver and kidneys progress.
Digital twins are virtual versions of organs that let researchers run simulated experiments on a computer rather than relying only on laboratory testing.
Professor Jon Chapman, head of the Mathematical Institute at the University of Oxford, said: “This exciting new partnership recognises the pioneering role that the Wolfson Centre for Mathematical Biology has played, and continues to play, in applying mathematics to understand diseases and their response to treatment.”
Founded by GSK, Imperial College London and the University of Oxford, the centre is intended to provide a new UK hub for research in modelling-informed medicine.
The centre is backed by £11m from GSK, alongside expertise in mathematics, data science and experimentation from the founding partners.
The partners aim to bring together fragmented research in the field, train a new generation of research and development specialists, share models on an open-source basis and build collaborations with further partners.
GSK plans to incorporate models of organs into its drug development pipeline within five years, aided by industrial placements it will provide to researchers from the centre.
The programme is led by Professor Helen Byrne and Professor Philip Maini at the University of Oxford, Professor Steven Niederer at Imperial College London, and Dr Anna Sher at GSK.
Niederer said: “We have seen maths used for modelling aeroplanes and cars, and increasingly there is a realisation that this has benefits in biology, where you can perform virtual experiments in models of humans at great speed and a fraction of the usual cost.”
At Imperial, Niederer’s team will build patient-specific models of organs using artificial intelligence and biological datasets.
They will mathematically represent millions of cells in organs such as the lungs, and the cause-and-effect relationships they hold to one another, by modelling a proportion of cells found in the real organ.
Using these models, researchers could perform a simple laboratory experiment into the effect of a drug on a single lung cell and then simulate how this would translate into larger effects such as changes in airway behaviour.
The approach differs from computational methods that look for statistical patterns in biological data because these models represent cause and effect, which could make them more explainable and robust.
Eventually, the approach could allow clinicians to use digital twins of specific patients to tailor treatments in real time, an approach Niederer’s group is already testing with cardiac patients.
At the University of Oxford, experts will develop and apply models grounded in physics, physiology and pharmacology to advance understanding of disease processes and help design more effective treatments.
These will include multi-scale models that integrate molecular, cellular and organ-level processes with whole-body physiology.
They will use digital twins and virtual patients to simulate treatment responses, optimise dosing strategies and design computer-based clinical trials.
They will also contribute open-source tools, standards for reproducibility and case studies that show the impact of model-informed drug development.
MiMeC will focus on adopting a mathematical modelling-first mindset in the development of new therapies.
Dr Anna Sher is MiMeC co-director and quantitative systems pharmacology lead in the respiratory, immunology and inflammation research unit at GSK.
Sher said: “By cycling between computer modelling, learning from the results, making predictions and then testing them, we can make faster, better decisions in developing new medicines.
“The tools and models developed through MiMeC strengthen GSK’s ability to generate virtual patients and digital twins to run computer-based (in silico) clinical trials, analyse different data types, and test scientific ideas more efficiently.”
