2nd Workshop on Medical Systems Biology
28 Nov – 2 Dec 2011
The University of Manchester
An Overview
In late November of 2011 a dedicated group of scientists – young and established academics, medical doctors and pharmaceutical professionals – gathered at the University of Manchester to continue the creation of a new field. The five day workshop “The Dynamics of Disease” looked into the application of “Systems Approaches” into areas of medicine and pharmacology.
Over the last decade, Systems Biology found its way into academia as the approach that tries to convert biology into a hard science. It is widely accepted by now that the analysis of biological data is no longer a matter of simple yes-or-no questions (“Is protein X present?”) and that understanding the response of a living organism to an intervention is often not just a matter of verbal description. Mathematical tools and particularly computational approaches based are becoming a “sine qua non” – no more biology without them. For a number of years it was obvious that if this Ansatz works it would also have impact on medicine. However, the step from academic biology to pathophysiology and clinical questions is neither simple nor is it quick. At that point the Workshop set its agenda: how can a widening of medicine (to become Systems Medicine) be achieved?
“The Dynamics of Disease” had lectures and tutorials to offer new research results and demonstrate how Systems Approaches work in practice. On the first day, Mike White argued that a mathematical description does much more than just confirm experimental findings. The process of building up a set of rules to computationally study the dynamics of a living system is a major driver in asking new experimental questions. To some Systems Biology with its stress on the “Omics” has become a synonym of “measuring everything”. This however, is a grave misconception. It is close to nonsensical to fill data bases without a prior concept of what model the data will contribute to. Experimental studies must be guided by computational investigations of a mathematical model to become truly valuable. The mental models we got so used to in the 20th century are no successively being incorporated into mathematical models of successively higher complexity. Therefore, we need experimental and clinical data that address the needs of the existing models.
Throughout the workshop this became a recurrent theme. In order to understand the details of disease dynamics, new ways of looking at the phenomena are required. And the most reliable source of reasoning about “what to look for” came from mathematical modelling. In some cases, disease patterns have periodic structure. Then the need for time-resolved measurements is clear. But if there is no temporal structure at the macroscopic level, this does not mean that there is no underlying temporal structure at. e.g. the cellular level. Whether or not a cellular rhythm scales up to the level of organismic physiology is a question of the number of cells that participate and how their behavior is coordinated. Synchrony amplifies, desynchrony blurs.
The question of optimal spatial and temporal markers of a disease to capture the most relevant events also highlighted that the analysis and modelling techniques also need further development. The other recurring theme was that current modelling is too much focussed on individual scales. Multi-scale models are the obvious answer but in only a few exceptional cases has this already been successfully achieved. Still, physiological models mostly ignore cellular and molecular knowledge. Still, cellular and molecular models do not properly scale to the level of physiology. Anatomical models ignore temporal evolution, temporal models ignore anatomy. Yet a proper description of most diseases will need to include either level and component.
One matter of debate throughout the workshop was: do we need to build large-scale models which include as much detail as possible OR shall we work with sets of simplified models that just address one important aspect of a disease a time successfully. The answer to this question will not be absolute but depend on the motivation. Ideally, understanding is achieved when as much as possible is put together at the highest level of detail. Practically, successful application (diagnosis and therapy) is the ultimate benefit of research into disease. In the extremes the one leads to the in silico human, the other is just a black box spitting out correct answers. Research in both the academic and the engineering direction will continue.
Gerold Baier, 13 January 2012
