A concise history of Systems Biology |
|
|
Nicolas Gambardella [Initially written in 2009] To understand the emergence and rise of Systems Biology, I would like to focus on three very different timescales: the rise of the notion of systems in science, the rise of Systems Biology as a bona fide domain of the life sciences, and the establishment of Systems Biology as the new biology. Emergence of the notion of systems in scienceI am not a historian and realise the following is quite flaky. This is more an opinion than a knowledge of facts. So, the best I can hope for is that you, the reader, will take home this opinion and perhaps transform it into a historical reflection. From the 16th to 19th century (in physics) and 20th (in the life sciences), scientists were interested in describing the world as it was. These descriptions took place at the same level as the object described. This era sees the reign of classical mechanics, anatomy, chemical kinetics, physiology, etc. One designs “laws” that govern characteristics, such as the laws of movement and the allometric laws for embryonic development. In biology, this is the era of the “grandes fonctions”. Here, we touch on physiology as a paradigm of scientific investigation, independently of its etymology, etc. In physiology, an organ or a process was considered a black box. A whole series of inputs were applied, a whole series of outputs were recorded, and people tentatively designed a mathematical function that produced the correct output as a function of the input. The main flaw of this approach was its assumption that the set of inputs used to design the transformation function was sufficiently diverse to cover the ensemble of all possible outputs. No emerging behaviours were possible. From the 17th century onward, another way of probing the world appeared, which flourished until the 2nd world War in physics and the end of the 20th century for the life sciences, that is, reductionism. The idea was now to break things. To explain the characteristics, the behaviours, at level N, people observed the entities or phenomena of level N-1. This is the great era of statistical physics, quantum mechanics, biochemistry, structural biology, molecular biology, etc. Although enormously productive, this approach crashed at the contact of real-world and technology. Indeed, the behaviour of the isolated component is not the behaviour of this component once it is placed within a system of interacting components. Engineers realised that very early for mechanics, but in the life sciences at large this would take until the beginning of the 21th century to reach that point. Finally, from the beginning of the 20th century, the notion of systems appeared. This is far larger than biology, of course. Here, to describe a system at level N, one puts together interacting entities of level N-1. The crucial concept is “put together”. This is not physiology, because one does not know – or impose – beforehand what will be the resulting, emergent, properties. This is the era of cybernetics (Wiener, 1948), information theory (Shannon, 1948), and, finally, Systems Biology. One often quotes Ludwig Von Bertalanffy for Systems Biology (Bertalanffy, 1968), but this is not completely true. LVB is one of the fathers of General Systems Theory, which applies more widely (although he was apparently the first person to coin the term Systems Biology in 1928. However, reading LVB really shows how general the systems approach is. Rise of Systems Biology as a paradigm in life sciencePhysiology was, of course, tremendously productive. From observation, physiology moved to experimentation, with the likes of Claude Bernard in the mid-19th century. Until the mid-20th century, physiology remained the dominant paradigm in all biological disciplines, with people like Pavlov and Langley, but also Hill and Meyerhof when it comes to the molecular level. And, although progressively being replaced by reductionism, it still survived for a relatively long time in the neurosciences, for instance, see Eccles, Katz (molecular neuroscience only made it in France in the 1980s, one of the longest battles of my PhD mentor, Jean-Pierre Changeux). From the beginning of the 20th century, reductionism appeared in Biology, in particular in Molecular Biology, see Michaelis-Menten, Kossel for the proteins, Avery for DNA, etc. It became the new paradigm in the life sciences with the generations of Pauling, Watson/Crick, Monod/Jacob. And here is an interesting parallel with Systems Biology: retrospectively, the work of all those scientists belongs to Molecular Biology. But it was not always called like that. When I started university, at the end of the 1980s, there were still debates on what actually constituted Molecular Biology, many people equating it to “recombinant DNA”. That despite the fact that the “Journal of Molecular Biology” was created in 1960 and published primarily structural biology papers. Biochemistry, structural biology, DNA recombination, all those approaches are molecular biology in the grand scheme of things, i.e, analysing dissociated living systems in test tubes. And this is an important point. We must differentiate between the techniques and the scientific paradigms. This will become important for defining Systems Biology. Finally, from the mid-20th century, another way of studying living systems appears. Personally, I pinpoint the real birth of Systems Biology to the papers of Hodgkin and Huxley in 1952. For the first time, someone studied different components (sodium and potassium channels), designed models for them, using the classical black-box approach of physiology applied to experimental measurement obtained by reductionist methods, put these models together in a bigger one, and simulated the resulting model (with a hand-driven digital calculator!) to understand how it worked. And the result, 58 years afterwards, still stands as an extraordinary success. In parallel, Britton Chance built electronic analogues of biochemical circuits to model and analyse enzymatic action. Funny enough, the very same year, Alan Turing, in his famous modelling paper on morphogens, suggested: “One would like to be able to follow this more general process [the morphogen-driven patterning] mathematically also. The difficulties are, however, such that one cannot hope to have any very embracing theory of such processes, beyond the statement of the equations. It might be possible, however, to treat a few particular cases in detail with the aid of a digital computer. This method has the advantage that it is not so necessary to make simplifying assumptions as it is when doing a more theoretical type of analysis”. Soon, other people followed, such as Denis Noble with the heart pacemaker. The 1960s were the period when systems theorists called for the formation of a discipline of Systems Biology. I already quoted Von Bertalanffy. We should also mention Mihajlo Mesarovic, who created a series of conferences and wrote books on the topic (Mesarovic, 1968). But it was too early. In particular, the data needed to turn the ideas into insights was not available. In parallel, what would become Computational Systems Biology was growing in two different fields: in Neuroscience, the cable approximation, (rev in Le Novère, 2007) became the standard way of modelling realistic neurons – with this approach, the electrical behaviour of a neuron emerges from the description of the behaviour of each of its segments – and in Biochemistry, for instance with the appearance of Biochemical Systems Theory (Savageau, 1969) and Metabolic Control Analysis (Kacser & Burns; Heinrich & Rapoport, 1974). The basis of what would become logic modelling was also laid out independently by Stuart Kauffman and René Thomas. Again, it was too early, and mathematical modelling of biochemical processes remained niche during the 1970s. Moreover, those efforts were literally wiped out by the tsunami of DNA recombinant successes. Modelling still demonstrated successes in the 1980s, for instance the Goldbeter-Koshland model of phosphorylation cascades, nine years before the experimental demonstration of the MAP kinase one. But the modelling of genetic, metabolic, and signalling pathways really took off in the 1990s, on model systems such as cell cycle (Tyson, 1991), bacterial chemotaxis (Bray et al, 1993), or MAP kinases (Huang & Ferrell, 1996), including the first model of “entire” metabolic networks (Joshi & Palsson, 1989). At the end of last century, the field was mature, with many scientists worldwide, a versatile modelling and simulation toolkit, and the ability to gather large amounts of quantitative data to parametrise the models. Establishment of Systems Biology as the new biologyFirst of all, “Systems Biology”, as the modern discipline (not as the philosophy for life sciences) – however you bend it – does not really originate from “omics”. Omics are about lists of components, while Systems Biology is about interactions. Functional genomics, as transcriptomics and proteomics (metabolomics came later), and Systems Biology were two different fields, almost entirely disconnected until 2004, when some visionary people from genomics realised that systems biology had made it and stopped fighting it systematically in the air and on the beaches. Instead, and rather than changing their science, they just rebranded themselves. Because they outnumbered systems biologists 50 to 1, in about a year they were able to rewrite history via reviews, grant applications and workshops. (For a good rant about it, I recommend reading the preface to Essays in Biochemistry - Systems Biology (Wolkenhauer et al, 2008). So, where does
modern Systems Biology come from? As a named discipline, it has two
different origins and two fathers. Originating from robotics, Hiroaki
Kitano spearheaded the “modelling systems biology” school
(Kitano, 2000). In 1998, he created the In parallel to
Kitano's efforts, Leroy Hood set up his own During a decade, the two communities ignored each other (sometimes in the primary sense of the word, they literally did not know there was another community calling themselves systems biologists and doing completely different things). But there are now (2009) signs of merging. In June, I attended a Nobel Symposium where the leading scientists of both sides were present, together with two of the main application fields, synthetic biology and cell reprogramming. [Update: in 2015, the president of the International Society for Computational Biology asked me to help bridging the gap, and I created the Community of Special Interest SysMod, which organise now a meeting at each ISMB]. Between 2000 and
2002, the number of articles related to Systems Biology increased
steadily in “regular” scientific journals, and the first
textbooks started to appear. To the extent that by 2003, when I
set-up my own research group, I was advised not to name it
Computational Systems Biology in case it was seen as just surfing on
a hype. Between 2000 and 2005, many large Systems Biology projects
were launched worldwide, amounting to several hundred million dollars
(e.g., the Alliance for Cellular Signalling, HepatoSys, the Yeast
Systems Biology Network), and In 2007, the BBSRC stopped this effort. Positive discrimination was no longer necessary to fund integrative approaches and hybrid experimental/modelling projects. Such projects were funded through regular calls, relevant to the existing domains of biology, development immunology, neurosciences, oncology, etc. Systems Biology was the new Biology, and Biology had join Physics, in being predictive and not solely descriptive and explanatory. References
|