Europhys. Lett., 68 (6), pp. 901–907 (2004) DOI: 10.1209/epl/i2004-10278-2 EUROPHYSICS LETTERS 15 December 2004 Genetic information and self-organized criticality P. R. Wills 1 ( ∗ ), J. M. Marshall 1,2 and P. J. Smith 1 1 Department of Physics, University of Auckland Private Bag 92019, Auckland, New Zealand 2 Department of Biomathematics, UCLA - Los Angeles, CA 90095-1766 USA received 23 August 2004; accepted in final form 14 October 2004 published online 17 November 2004 PACS. 87.23.Kg – Dynamics of evolution. PACS. 05.65.+b – Self-organized systems. PACS. 05.70.Jk – Critical point phenomena. Abstract. – The numerical fitnesses of species defined in the Bak-Sneppen model of self- organized criticality are interpreted as binary strings. This allows new species to be generated by mutation of survivors. It is shown that selection in Bak-Sneppen systems defined on both uniform and random lattices produces genotypes in conformity with the Eigen criterion for the accumulation of genetic information in macromolecular sequences. The Bak-Sneppen (BS) model [1] of self-organized criticality (SOC) mimics certain gross features of global evolutionary dynamics. It has been used to explain the frequency distri- bution of extinction events of widely varying magnitude deduced from the fossil record [2] and other biological phenomena of apparently fractal origin [3]. The model has been applied to mutation in bacterial populations [4, 5], but little effort has been made explicitly to take account of phenomena that arise directly from biological mechanisms. Here we consider a variant of SOC in which species are generated through genetic replica- tion (inheritance and mutation) and the selective fitness of a species is a phenotypic property. We find that the BS model conforms to the Eigen criterion [6,7] for the accumulation of genetic sequence information. This establishes for the first time a direct connection between disparate approaches to studying evolution: extremal dynamics on lattices of variable dimension; and population dynamics on sequence-encoded fitness landscapes. The basic one-dimensional BS model consists of a linear array of cells, each of which contains a species that has an assigned fitness (or “barrier”) B, selected from a uniform dis- tribution, typically in the range [0 1). At each time-step the species with the lowest fitness is eliminated, along with its lattice neighbours, and each is replaced by a new species with a ran- domly assigned fitness value. The system evolves eventually to a critical state characterized by a threshold value of the fitness B c which, in the thermodynamic limit, determines the ultimate fate of any species: death or survival. The model has been studied in high-dimensional sys- tems [8,9] and applied to the study of random [10], branched [11] and scale-free [12] networks. ( * ) E-mail: p.wills@auckland.ac.nz c  EDP Sciences