Landscapes and Molecular Evolution

Evolution of RNA molecules in vitro is visualized as a hill-climbing process on a fitness landscape that can be derived from molecular properties and functions. The optimization process is shaped by a high degree of redundance in sequence-to-structure mappings: there are many more sequences than structures, and sequences folding into the same structure are (almost) randomly distributed in sequence space. Two consequences of this redundance are important for evolution: shape space covering by small connected regions in sequence space and the existence of extended neutral networks. Both results together explain how nature can fast and efficiently find solutions to complex optimization problems by trial and error while the number of possible genotypes exceeds all imagination. In the presence of neutral networks, populations avoid being caught in evolutionary traps and eventually reach the global optimum through a composite dynamics of adaptive walks and random drift. Results derived from mathematical analysis are confronted with the results of computer simulation and available experimental data.