Sijbren Otto and Fred Ludlow at the University of Cambridge, UK, call on chemists to embrace complexity and take up the emerging discipline of systems chemistry

Complex systems are everywhere, from stock markets and the web to ecosystems and metabolic pathways. The fields of biological and atmospheric chemistry deal with complex systems out of necessity, but only recently have synthetic chemical networks begun to appear. Modern analytical techniques have turned systems, which would once have been an intractable mess, into a mine of data which, with the right tools, can be converted into new knowledge. 

Systems chemistry deals with the emergent properties of interacting chemical systems or networks. In other words, properties that result from the interaction between the components in a network, rather than any one species acting individually. The networks can be broadly divided into two types, those under thermodynamic control, and those under kinetic control. One example which operates under thermodynamic control is the dynamic combinatorial library (DCL). In a DCL, a network of reversibly linked building blocks is allowed to equilibrate, resulting in a product distribution that depends on the relative stabilities of the possible oligomers. Under thermodynamic control, the product distribution can be altered by preferentially stabilising one oligomer. This can be exploited for the identification of new host-guest systems - those oligomers which are amplified by a guest are likely to bind strongly to it.

In addition, quantitative information about the microscopic properties of the constituents (host-guest affinities) can be determined from the macroscopic properties (concentrations) of the whole system. By measuring the concentration of each oligomer in the network under a variety of initial conditions (building block concentrations), a quantitative model of the library can be built and fitted, thus quantifying the various host-guest interactions within the system.

For networks operating under kinetic control the history of the system is important, and this creates the possibility of emergent temporal and spatial patterns. One example of this, and a favourite at chemistry open days, is the Belousov-Zhabotinsky reaction. When stirred, the concentrations of various intermediates oscillate periodically, but when left unstirred, moving chemical fronts (waves) can be seen. These waves are not the property of any individual species in the solution, but instead of the interlinked and auto-catalytic reactions occurring simultaneously.

One area of particular interest is the study of self-replicating molecules. The networks of cross and autocatalytic reactions involved such systems provide chemical models for the primordial chemical networks that eventually became living organisms. Pioneering work includes a study by Ghadiri and co-workers on a network that arises from a set of nine self-replicating peptides and their pre-cursors. Despite its small size, the network exhibited features usually associated with larger networks, such as a hierarchical topology. The behaviour of sub-networks in isolation also differed from their behaviour within a larger network, with different pathways operating under different starting conditions. 

The ability of synthetic chemical networks to mimic some of the features of much more complex, but fragile, biological networks makes them ideal model systems, from which we can hope to gain some insight into the common organisational principles behind a range of complex networks. This may in turn lead to a better understanding of how to modify biological systems effectively, engineer more complicated functional systems, or even provide us with clues to the origin of life.

Read Sijbren Otto and Fred Ludlow's Tutorial Review on 'Systems chemistry' in issue 1, 2008 of  Chem. Soc. Rev.