Seeing double: the rise of digital twins in life sciences

Little more than a decade ago, the concept of digital twins in life science research may have sounded more like science fiction than a viable model for research and development (R&D). However, with COVID-19 rapidly accelerating the adoption of digital technologies, the industry is warming to the concept of digitally aided biology.

Digital twins are a prominent figure in this movement, having already made their mark improving operations in the manufacturing sector.

So, what exactly is a digital twin? According to IBM, the technology allows users to develop an exact virtual replica of a physical entity or process. Connected sensors collect data from the physical asset that can be incorporated into the digital model. In other industries, this could mean creating a complex digital model of an electricity grid or spacecraft and using data to inform test scenarios to identify potential issues.

In life sciences, the use of digital twins is still in its infancy. However, as digitisation takes centre stage, companies are beginning to warm up to innovative new concepts.

“I think a lot of people can see the potential of digital twins,” says chief scientific officer and co-founder of Synthace, Markus Gershater. “That possibility for us to be running better and more informative experiments that are transferable, from experiment to experiment and from experiment into the clinic.”

For Gershater, the use of digital twins in life sciences research can be separated into two categories: biological and experimental. While the likelihood of realising a full biological model of a cell or human is still a long way away, experimental digital twins have the potential to revolutionise clinical research now.

Digitising the lab environment

It’s no secret that life sciences are undergoing a transformative push to adopt digital technologies. For years, researchers have done their best to accommodate the evolving complexity of their field with the limited time and resources available to them, even as emerging areas, such as genomics and personalised medicine, significantly increased the process of discovery and development.

“There’s a real mismatch between the complexity of what we’re trying to do and the tools we have available,” explains Gershater. “The way we currently work is mostly manual, and it often relies on quite simple experimental design and analysis. This is the kind of analysis that you’d be able to do in Excel and tends to be the common level of analysis across biology.”

While much focus has been dedicated to using digital twins in manufacturing and modelling, for Gershater, the technology has significant potential in the experimental space.

Issues with costs and resources often hamper research projects, as clinicians can only do so much with the tools they have at their disposal. As a result, when researchers embark on a new drug discovery project, the odds are largely stacked against them. Approximately 90% of drug discovery fails – a substantial amount given that the global pharma industry spent almost $200 billion on research and development in 2020.

“Physical experiments are hugely costly,” explains Gershater. “If we can reduce the number of experiments we can run and make them more targeted, make them more effective, and less likely to fail, then that’s a huge potential benefit.”

This is where digital twins can be a useful asset in R&D. Whereas it may take life science researchers months, even years, of dedicated focus to sort and analyse data, advances in computing mean that digital twins can run multiple test scenarios simultaneously. Moreover, automating testing allows clinicians to rapidly recreate and reproduce trial scenarios, often conducted in highly controlled environments, across locations and personnel.

“It offers the possibility for a separation of labour,” says Gershater. “Now you have people that can design and plan experiments, and then different people then run those and who are experts in the running of those experiments.”

Harnessing the power of data

Of course, powering these sophisticated systems requires vast amounts of data. While the complexity of human biology hasn’t changed, our ability to understand it has substantially improved thanks to progress in sensor technology, artificial intelligence and machine learning. Consequently, biologists have multiple avenues they can use to garner new information about the performance of an experiment.

High-quality historical data is far more complicated to source. One of the most significant issues facing life science researchers is that there is still a lot that we don’t know about human biology, which means there is no existing data to inform digital twin models in specific areas. This issue is compounded by the small and messy data sets commonly found in life science research.

“What we need to make progress in modelling such a complex system is a large amount of exceptionally high-quality and deep data that can give us a lot of insight into that system,” says Gershater. “A step to get there would be the experimental digital twin, which then aids in actually generating those complete and high-quality data sets.”

Incorporating digital twins into experiments early on can help to boost the volume of usable data. By designing and running scenarios in the digital domain, researchers can gather both molecular data and metadata detailing the purpose and process of an experiment in a coherent and structured set.

Given enough historical, real-time and process data, researchers can leverage the machine-learning capabilities to understand not only how a target is performing, but how it will behave in the future. Ultimately, the ability to predict how certain molecules and drug targets will behave in the real world could prove to be a crucial element in pursuing personalised medicine.

Early iterations of digital organ models, such as Siemens Healthineers effort to develop a digital twin of the heart and Dassault Systems living brain project, showcase just how revolutionary these concepts could be if widely applied to medical treatment. Suppose clinicians could create exact digital replicas of individual patients, using real-time data to track performance and inform decisions. In that case, they could tailor each treatment to optimise the likelihood of success.

As Gershater explains, “The ideal system will be where we built up hugely sophisticated structured datasets, which have informed the production of true biological digital twins of cells of disease processes of tissues of human beings.”

Entering the metaverse

Driven by the need to accelerate and improve the success of clinical research and drug discovery, life science companies are increasingly looking to technology to plug the gaps between physical lab work and complex data analytics.

Creating virtual replicas using digital twins is just one small part of the ongoing digital transformation. The idea of immersive computing experiences has already caught the attention of investors across the industry. Now with the buzz around digital still going strong, the hype around the ‘metaverse’ has reached healthcare as companies seek ways to improve processes

“Computer aided biology is essentially a vision of what would be possible if instead of these manual methods and more simple methods, we applied cutting-edge digital and automation technologies to our biology R&D,” says Gershater. “That would be experiments and processes that can be designed and planned in the cloud, with digital help in designing and planning. Experiments can then be run in highly automated labs so that our ability to conduct research is augmented by technology as well.”

Although metaverse technologies have not reached the mainstream consciousness of life science companies, according to analysis from Rock Health, there is growing evidence that investors are beginning to throw their weight behind immersive digital technologies. As the report notes, “2021’s investors shelled out $198 million in funding for US digital health start-ups integrating VR or AR technologies across 11 deals, more than double the $93 million raised across eight deals in 2020.”

In theory, a healthcare metaverse would combine the best elements of modern digital systems and traditional physical interactions. For example, using a virtual reality headset to practise surgical techniques using anatomical holograms – the physical training remains the same, but the virtual system allows for multiple attempts without creating waste.

“There are much more effective ways of working with biology. You don’t even need to reach the giddy heights of supercomputing in AI; just running complex, sophisticated experiments can truly give you an insight into the complexities of biology,” explains Gershater. “These technologies and capabilities have been around for a very long time and yet they haven’t been adopted by biology.”

A future of personalised medicine

While the promise of data-driven research and medical treatment sounds like a promising area for innovation, it remains a hugely costly venture. The challenges associated with successfully developing digital simulations of these complex biological entities has been likened to the Human Genome Project (HGP).

It is not an unfair comparison, as both projects are highly ambitious undertakings. But, just as the HGP advanced our understanding of the human genome to 99.9%, digital twins can dramatically overhaul our abilities in discovery, research, and ultimately the way we treat patients.

In an ideal world, digital twins would already be streamlining research projects worldwide. But unfortunately, the reality is far more complex than that. For Gershater, the process is more likely to comprise a collection of smaller steps, each building upon the success of the last until key decision-makers in the industry are convinced that digital twins and automation technology can live up to the hype – as well as the expense.

“It’s the ultimate combination of personalised therapy,” he says. “An experimental digital twin will be a massive leap forward. We’ve seen what happens where you have the basic rudimentary implementation. It changes the art of the possible quite dramatically, which is a massive step-change in our understanding of biology, human physiology and disease.”

About the author

Eloise McLennan is the editor for pharmaphorum’s Deep Dive magazine. She has been a journalist and editor in the healthcare field for more than five years and has worked at several leading publications in the UK.

Don’t miss your complimentary subscription to Deep Dive and our newsletter

Sign up



Your name

Your e-mail

Name receiver

E-mail address receiver

Your message