Towards a Philosophy of Systems Biology

 

Hans Westerhoff, Fred Boogerd, Frank Bruggeman

Vrije Universiteit, The Netherlands

and Jan-Hendrik Hofmeyr

University of Stellenbosch, South Africa

 

The face of Biology has been changing. Not only has it become the most rapidly advancing natural science, it has also evolved into 'Big Science', requiring investments such as for the sequencing of the human genome. As such, Biology is comparable to Physics at the time of the major advances in particle physics and the construction of the first particle accelerators. Biology has also changed feet however, feet in the sense of philosophical and methodological foundations. The latter change is addressed in this symposium.

The objects of study for Biology are living systems. To be alive autonomously, a system requires more than 300 processes, catalyzed by proteins and encoded in nucleic acids, at some 1000 basepairs per gene. In nonlinear interactions new properties emerge that are important for the maintenance of the living state. If only for its sustenance any of the 300(+) processes depends on all the other 299(+) processes. Consequently (i) the living state cannot be understood by a scientific strategy that studies 300/n times n well-defined components in isolation, (ii) when studying part of the living system without following the behavior of all 300+ components, the verification or falsification of an hypothesis cannot be definitive, and (iii) when studying the entire system without the determination of how those 300+ components are causally interconnected and parametrised to account for their change Biology will only describe and not explain and ultimately understand Life. Hence 'proper' Biology can never meet the criteria for the 'proper' science defined by the physicists and philosophers for whom Biology was only 'stamp collecting'. Biology cannot achieve the ideal of 'Physics' as put forward by Rutherford (which we shall here refer to as 'Physics').

But then, Molecular Biology flourished. It led to the characterization of the structure and mechanism of action of many molecular constituents of life, to the mapping of pathways, and ultimately to the ability to sequence every single gene and measure every single mRNA molecule. An exercise of which no clear example exists in Physics. Although Molecular Biology started from the perspective of applying the 'Physics methodology' to Biology, it became most successful when it stopped doing that strictly: By refraining from the precise quantitative testing of all detailed predictions, Molecular Biology was able to generate a cartoon type of understanding of many of the pathways operating in organisms. Hard testing of the functioning of molecules and pathways in the living state was impossible, because not all other molecules had been identified and the above experimental means were not available. Therewith, Molecular Biology has become another instance of a successful science that does not really follow the paradigms of the 'proper' science Physics. However, the reason for its success was at the same time a weakness: it lacks predictive power of the outcome of precise changes made to those cartoons and Molecular Biology did not address Life itself, i.e. living systems in their entirety. As such it was not a substitute for Biology at large.

 

Now (or soon) however, all molecules of complete living systems (such as yeast or E. coli) can be identified (or will become so soon) and manipulated. Accordingly, the limitations that kept Biology from becoming 'proper' science have been removed. Entire systems can now be studied experimentally in completely defined ways. Will, or should, Biology now finally become Physics rather than stamp collecting?

Engineers may think it should. Principles) that suit man-made systems (such as robustness or efficiency) are now recognized in (or projected onto) biological systems. Physicists and statisticians may say it should, and analyze the expression patterns using singular value decomposition inspired by linear systems. Molecular Biologists may say it would by becoming 300 times faster Molecular Biology.

 

We doubt whether in terms of its methodology and philosophical foundations Biology should become Physics. The new Biology cannot pretend that Life can be understood by studying 300 single molecules separately. There will be no two engineering principles that explain Life. Biology does need to delve into the living cells, connect to the molecules and find regularities in the mechanisms at work. The new Biology should focus on what emerges from the interactions. It must come to understand the networks of precisely interacting molecules. Simplicity of explanations cannot be an aim as it is in so many branches of Physics. Biology should have functional concepts, but these should transcend those of engineering and physics (such as robustness and phase transitions) and deal with purely biological phenomena such as physiological and genetic adaptation. The plasticity of adaptation may lead to biological laws akin to yet entirely different from the celebrated laws in Physics such as the second law of thermodynamics. The commonalities between different cell types may serve as the generalities required for theories, rather than that cosmic universalities derivable from underlying physical principles alone, should do so. Biological systems are subject to challenges that are entirely different from those of the man-made systems studied in Engineering and the inanimate systems studied in Physics. Accordingly, we should expect to discover evolved molecular mechanisms that have no resemblance with any known mechanism studied in engineering nor physics. Biological systems are heterogeneous; the ensemble averages of independent identical components that lead to universal laws in Physics, may not pertain to Biology; the laws in Biology may not be based on the same principles altogether. Yet, Biological Science may achieve similar generality owing to principles of natural selection of traits vis-à-vis complex challenges, limited possibilities, and common ancestors. The new Biology should also have theories that explain qualitatively new functioning on the basis of quantitative interactions. Whereas in Physics many phenomena can be qualitatively explained without raising suspicion (e.g. magnetization of a magnet in the 2D Ising model), validation of theories in the new Biology are anticipated to depend highly on quantitative testing, e.g. because of parallel regulatory mechanisms. Correlations among system properties will not suffice. It is the spiraling causality at work through molecular mechanisms in cells, that the new Biology should go for.

 

We associated much of the existing philosophy of 'proper science' with that of 'proper' Physics (in the above sense). We suspect that there may be different types of science that are equally 'proper'. The new Biology, which is often called Systems Biology, may be a youngster among these: To us it appears that Systems Biology should now be enabled to stand on its own philosophical and methodological feet. These feet should relate to philosophical underpinnings, to methodology, to theory/model/simulation relations, as well as to the importance of organizational principles vis-à-vis physical laws.

We hope that this symposium will help put Systems Biology on such feet, i.e. provide it with proper philosophical and methodological foundations.