Modeling and simulation have never really been a core component of research for mainstream biologists in the way that they are for scientists working in physics for example. With biology still in its adolescence as a science by comparison with other fields and its researchers still learning how to deal with the incredible complexity of the systems that they study, it is perhaps not surprising that biologists have been relative latecomers to the modeling game.
Capturing the trajectory of a satellite in lines of computer code is a much more tractable problem than simulating on a computer, even the simplest and most well-studied of biological organisms. At the molecular level of course, biology is essentially physics and chemistry, but at any kind of scale at which the term “biological” can be applied, you’re being forced to deal with an enormous number of moving parts, significant heterogeneity between one copy of the system and the next and a degree of stochasticity that makes you wonder how organisms ever get anything done, let alone survive and thrive in hostile environments. In a word – biology is messy.
Scientists of all persuasions have always been modelers, whether or not they recognize this fact or would actually apply the label to themselves. All scientific concepts are essentially models since they are a description of things and not the things themselves – and the advancement of science has been largely founded upon of the relentless testing and improvement of these models and their rejection in the case where they fail as consistent descriptions of the world that we observe through experimentation.
Many (arguably all) of the models that are currently accepted by the scientific community are incomplete to some degree or other, but even an incomplete model may often have great value as the best description that we have to date, of the phenomena that it describes. Scientists have also learned to accept the incomplete and transient nature of such models since it is recognized that they provide a foundation upon which more accurate or even radically new (and hopefully better) models can be arrived at through the diligent application of the scientific method.
As technological advances have broadened the scope and refined the detail of biologists’ observations of the cell, it has forced them to confront as never before, a degree of complexity that defies the ability of any single researcher to maintain an accurate mental model of all of the moving parts, let alone to understand the complex behavior that results from their combinatorial expression. It is precisely under such circumstances that models can be invaluable as vehicles to organize and orient knowledge and to provide intellectual frameworks within which to reason about these complex collections of interdependent parts.
If your modeling system is suitably transparent and flexible, even a partial model can be of great value since the missing pieces of the model are a fertile breeding ground for new hypotheses and the model itself can even provide a framework for testing them. Physicists understand this paradigm well. For example, the inability of astronomers to be able to map the positions and motions of a particular binary star system to the Newtonian gravitational model can actually provide the means to discover new planets – the divergence of the observations from the gravitational model being a measure of the amount of mass unaccounted for and its position in the sky. Models therefore, can clearly have “predictive” value, even when they are “wrong”.
This is indeed good news for biologists to whom computers now begin to afford the possibility of modeling the events that occur within a living cell. If we consider signaling pathways for example, while it is true that the catalog of signaling components and their interactions is growing at an accelerating pace, it is still far from complete. In the pathways for which we think we have a pretty good idea of who all of the players are and who interacts with whom, we can lay all of this knowledge out in a pretty pathway map but just like a street map, it does not give us any idea of how the traffic flows through the city. The map of the pathway is itself a kind of static model in which the essential components of causality and time are absent – a useful first step, but still not the kind of living, breathing model that we need to capture the behavior of living, breathing biological systems.
Modeling can be an invaluable adjunct to experimental research, in some cases even a replacement for some part of it. The ability of physicists to accurately simulate the explosion of a nuclear weapon on a computer has largely obviated the necessity for military nations to test nuclear weapons and civil engineers routinely test the load-bearing capacity of new structures using computers, before these structures are ever built or even mocked-up. But while biological researchers may still be far from ever being able to hang up their lab coats and retire their pipettes, there seems little doubt that as the computational biologists learn to more accurately reproduce the complex causal and temporal patterns of events that occur in living cells, researchers in this young field will come to embrace virtual experiments alongside their lab work, just as their peers in other fields have, hopefully bringing to biology, all of the manifold benefits that the modeling approach has brought to other fields.
© The Digital Biologist