SIMULATING AN EXPLORATION OF RNA CONFORMATION SPACE WITH AN APPROPRIATE PARALLEL-UPDATING STRATEGY

A thermodynamic ensemble of stable RNA structures should emerge in the long-time limit as a result of an expeditious exploration of conformation space. The design of a simulation of this exploration giving consistent results in the long-time limit has been hindered by two main factors: (1) the need to incorporate the kinetic or activation-energy barriers for structure conversion and (2) the possibility of competing folding pathways. In this work, we implement a parallel kinetically controlled simulation that encompasses both aspects and ultimately yields all significant contributors to the thermodynamic ensemble. We provide evidence supporting the conjecture that thermodynamic representativity may be reproduced by introducing an appropriate learning or updating strategy in a kinetically controlled parallel simulation. The results are specialized to the illustrative cases of a transfer RNA and a midivariant-1 Qbeta RNA for which the dominant native and non-native structures have been independently established. In both cases the thermodynamic ensemble is reproduced.