A stochastic theory of macromolecular evolution (Q2266682)

The stochastic theory of macromolecular evolution starts from a description of the transition probabilities of certain elementary events such as replication and destruction and extends the birth and death processes so indicated to more general branching processes through the inclusion of mutational transitions. The main result of the present paper is to establish that a local version of the error threshold, which is called a localization threshold, defines limits of metastability of macromolecular distributions around locally ''best'' sequences with an expression for the metastable lifetime. These metaspecies succeed one another in a fashion epitomizing the Darwinian optimization principle. On the other hand, beyond the localization threshold, ''survival of the survivor'' narrowing and ''random drift'' diffusion compete to provide a fluctuating dynamics in which some information is nevertheless dynamically preserved. The programme proves both natural and instructive to first consider a linear theory. In section 2, a hierarchy of equations for the probability distribution of the populations of the various mutation coupled sequences is formulated. In the same time, the asymptotic reduction to deterministic behaviour is established. Then in section 3, the stochastic theory to a finite population size is given and a non-linearly coupled hierarchy is formulated. It is established that, in the absence of mutations, fluctuations are small of the order of the square root of the total population once equilibrium is established. Finally, the phenomenon of ''random-drift'' is shown to become the predominant mechanism of change in a finite population as optimization approaches completion. In section 4, a summary of the implications of the stochastic theory of self-replicating macro- molecules is presented.