Approximations, idealizations and ‘experiments’ at the physics–biology interface

Abstract

This paper, which is based on recent empirical research at the University of Leeds, the University of Edinburgh, and the University of Bristol, presents two difficulties which arise when condensed matter physicists interact with molecular biologists: (1) the former use models which appear to be too coarse-grained, approximate and/or idealized to serve a useful scientific purpose to the latter; and (2) the latter have a rather narrower view of what counts as an experiment, particularly when it comes to computer simulations, than the former. It argues that these findings are related; that computer simulations are considered to be undeserving of experimental status, by molecular biologists, precisely because of the idealizations and approximations that they involve. The complexity of biological systems is a key factor. The paper concludes by critically examining whether the new research programme of ‘systems biology’ offers a genuine alternative to the modelling strategies used by physicists. It argues that it does not.

Introduction

As part of the project which led to this special issue, I conducted interviews with various physicists and biologists—mainly at the University of Leeds, but also at the University of Edinburgh and the University of Bristol—about the differences they perceive between the disciplines of physics and biology, and was also a ‘fly on the wall’ in situations in which physicists and biologists interacted. My interviews and observations focused on those that work at the interface of the two disciplines in soft condensed matter physics of biological systems, and particularly on the difficulties that they experience when working with colleagues ‘on the other side of the fence’ in molecular biology. To get a more rounded picture, I’ve also interviewed molecular biologists who are familiar with the work of, and in some cases have co-operated with, those soft condensed matter physicists.

The main aim of the study was to isolate differences with potential epistemic significance, especially in method and attitude, for philosophical examination. A qualitative approach was selected in order to enable an appropriate focus on detail, and to provide the scientists with maximum flexibility in presenting their views and accounts. The interviews did not involve set questions, but took the format of guided discussions, where the interviewees were encouraged to share their thoughts about the differences between condensed matter physics and molecular biology, and offer testimony in support.

The sample size was relatively small, so the result should be taken to frame tentative hypotheses which are worthy of subsequent quantitative study.¹ I should add, however, that my findings in this paper are also based on informal discussions with other scientists, on published work at the interface, and (to reiterate) what I have seen in group scenarios (including conferences). I also take them to be independently supported by Bueno (this issue), who argues that ‘molecular biology and physics seem to meet in a curious, rather disjoint, way at the nanoscale’, and by what Calvert and Fujimura (this issue) say about systems biology (as well as their recognition that ‘reductionism and holism are “fighting words” in the history of biology’). At the very least, my findings illustrate interesting differences that sometimes occur at the aforementioned interface; and considering these is worthwhile even if they are not as widespread as I believe—by way of a somewhat bold hypothesis—that they are.

I will start by presenting two episodes which reflect differences in approach between those working in the two specific sub-disciplines, and I will use data from each camp in order to portray these as accurately as possible. Both illustrate my first claim that molecular biologists have a strong dislike for models involving a high degree of idealization and approximation, and prepare the ground for the philosophical discussion in the remainder of the paper. But before I begin, there are two further points that I should like to emphasize.

First, we should only take these stories to reflect trends when it comes to considering differences at the disciplinary level, even if we put aside the issue of the small sample size, due to the considerable disunity within the disciplines themselves.² For example, one of my interviewees studies animal movement (e.g. by using classical mechanics) and was supervised by a physicist during his PhD despite now being based in a biology department. Somewhat unsurprisingly, I found that his attitudes on several key issues bore a strong resemblance to those of the soft condensed matter physicists. For example, he sees ‘the flow of ideas coming very predominantly from the more fundamental sciences’, is ‘a great believer in simple models’, and doesn’t ‘really believe in the divisions between the sciences … at the research level’ (Interview, July 2007).

Second, and on a closely related note, it should be remembered that chemistry is closely related to molecular biology (which I have even heard described as ‘applied biochemistry’), and therefore that I was actually investigating an area where three disciplines come together. One molecular biologist, Professor Alpha, told me ‘I would still describe myself as a chemist, even though the chemistry I’m looking at mostly involves non-covalent interactions whereas most card carrying chemists do things with covalent bonds.’ (Interview, January 2008). Another who also studied chemistry as an undergraduate, Professor Beta, explained that ‘There’s a real difference in my world between chemists and physicists … The chemists, when they want to, have little difficulty in mastering the biological problem … [and] they don’t mind doing boring chemistry in order to [help to solve it] … With physicists I’ve found much more of a distance. They—to my surprise … —have much more difficulty grasping the biological problem …’ (Interview, December 2007). It should be added that when it comes to computer simulations, in particular, one often finds highly similar work being performed in ‘computational biophysics’ and ‘computational chemistry’.³