Can all aspects of life be simulated in a computer or are there properties of living systems that in fact cannot be represented? How is this related to different types of complexity we observe in living systems that we have to deal with when modeling such systems? How does this influence the modeling process from purpose to predictions – and their use for decision-making in society?

In the research projects, mathematical and computational modeling is used quite extensively. This is not new in biology, in fact Vito Volterra, whose name we have used on our high profile lecture series, came from mathematics and physics and applied this knowledge to biology in the early 1900. Similarly, the well-known Michaelis-Menten equation from 1913, which represents mathematically the rate of (at least some) enzyme catalyzed reactions, was formulated before one knew what enzymes are.

So what is a formal model? In life science such models are representations of phenomena in biological systems in the form of mathematical equations or as a computer program. If it is simple enough, the model can in principle be “executed” on a piece of paper, however, in practical terms, most formal models will need a computer to run or to be analyzed.

So, how much complexity can be represented in a computer-executed model of let’s say a simple bacteria – can we model all aspects of it if we had sufficient computer power? Our invited lecturer Dominique Chu (School of Computing, University of Kent) took us through different aspects of complexity in relation to modeling of living systems. In the real world things just happen “by themselves”, however, a model will need to include everything that can happen, unless one can describe the model in terms of first principles and the processes that occur will emerge as a result of that. In a computer the software will not alter the hardware, this is not necessarily the case in living systems, where there also is no clear distinction between software and hardware.

This type of complexity was one of Robert Rosen´s key interest, in fact he claimed to prove that this could never happen – that a computer could never simulate all aspects of life – and this is in fact one of the key properties of living systems. He argued that living systems are special in being what he referred to as closed to efficient entailment - self sustained, self repaired, self replicating. One of the smallest such system he studied was the Metabolism-Replacement (M,R) system, where mappings of external sets of compounds into internal set of metabolites is performed by a set of transformations that must themselves be formed by the metabolites and a self-maintained transforming capacity (suggested reading below).

Still, although there are clear limits to what models can provide or predict, we can use formal models to learn more about living systems and the interplay between different components or processes. The modeling exercise then becomes very important. What is the purpose of the model? Which modeling approach should we choose to best fit this purpose and the data available or obtainable by our methods/technologies? Given these choices, what is the predictive power of the model and if so, to what degree can this be used for decision making at a societal level?

During the workshop these issues were discussed in general and for the different models/modeling approaches that were presented by participants attending the workshop. In fact, the participants were given the task of presenting challenges and reasons for why their model is limited in its predictive power in a pure scientific setting, but also in the setting of its extended application, typically the type formulated in grant applications. These questions very soon touches on responsibility aspects of the scientific process and on the use of formal models therein and in decision processes outside science. Roger Strand took the discussion very nicely in this context at the end of the workshop. A discussion that fits well within the scope of DLN, a center addressing both digital biotechnology and RRI.

Modeling living systems poses many challenges; on facing complexity issues, on the modeling exercise itself, and on the increased demand for providing a scientific basis behind decision-making and policies. These challenges are not discussed on a daily basis in the research projects and arenas where such issues can be lifted are probably needed. If so, DLN should provide an opportunity for doing this as one of many activities in the center and for scientists outside DLN.

*Some suggestions for further reading*

Life Itself. A comprehensive inquiry into the nature, origin, and fabrication of life. Robert Rosen, Columbia University Press.

Anticipatory Systems. Philosophical, Mathematical, and Methodological Foundations. Robert Rosen, Springer (2012).

Chu, Dominique (2011) Complexity: against systems, Theory in Biosciences 130, 229-245. And references therein.

The Rightful Place of Science: Science on the Verge (2016) Benessia, A., Funtowicz, S., Giampietro, M., Perira, A.G., Ravetz, J., Saltelli, A., Strand, R., and van der Sluijs, J. Consortium for Science, Policy and Outcomes, Tempe AZ and Washington, DC.

Publisert: 20. Sep 2017 - kl. 10:27

Sist oppdatert: 18. Oct 2017 - kl. 19:49