EFFold: A program for predicting a set of RNA secondary structures that are near optimal by perturbing the RNA folding, and simultaneously to compile all possible stems that are thermodynamically favorable in the RNA folding simulations.
We all know that although the rules of free energy paramreters in the RNA folding are derived from experimental measurements that have normal distributions of precision and accuracy, the free energy parameters are treated as precise in dynamic programming algorithm, such as mfold. The primary approach in EFFold is to simulate a normal distribution of the energy rules by perturbing the free energy parameters of Turner's energy rules within the range of experimental errors under the predetermined parameters. Thus, uncertainties of thermodynamic parameters for the formation of RNA duplexes and loops in Turner's energy rules used in the dynamic folding are reasonably considered in the method EFFold.
In practice we often generate 50 or 100 artificial ``simulated energy rules'' (SER) and then compute the corresponding 50 or 100 optimal structures with the lowest free energy by these artificial SER, respectively, using dynamic programming algorithm. These computed ``optimal'' structures are then compared and classified based on the structure similarity among them. A set of predicted structures can be ranked by means of their energy distribution computed from those optimal structures in the sample. The helical stems found to be thermodynamically favorable from the simulation are compiled. Those thermodynamically favorable stems can be used a pool of the structural element for constructing a phyloge- netic conserved structure. The program is based on the dynamic programming algorithm and implemented in Fortran 77 running on Unix.
Back to RNA Folding and Modeling at CCRNP
Back to CCRNP Research Projects