Selection in biochemical synthesis processes is a very important task: processes as protein or nucleic acid synthesis must be very accurate although the cell abounds with closely related molecules which, intuitively, should be difficult to distinguish. The mechanisms of selection and the underlying processes have been extensively studied in the last years, and the ideas have been developed into mathematical models to visualize the potentials of recognition and error correction. There are several aspects on this, the main aspect of the present work being the important fact that the cell must pay for the increased accuracy in terms of available free energy: processes are driven from equilibrium, and do away with appreciable dissipation.

Though the structures presented in crystallographic models of macromolecules appear to possess rock-like solidity, real proteins and nucleic acids are not particularly rigid. Most structural work to date has centred upon the native state of macromolecules, the most probable macromolecular form. But the native state of a molecule is merely its most abundant form, certainly not its only form. Thermodynamics requires that all other possible structural forms, however improbable, must also exist, albeit with representation corresponding to the factor exp( — Gi/RT) for each state of free energy Gi (see Moelwyn-Hughes, 1961), and one appreciates that each molecule within a population of molecules will in time explore the vast ensemble of possible structural states.