RNA inverse folding problem is a bioinformatics problem where the objective is to find an RNA sequence that folds into a given target secondary structure. In this work, we use Evolutionary Computation to solve a new and innovative multiobjective definition of this problem. In this new multiobjective definition of the problem, we have considered the similarity between target and predicted structures as a constraint, and three objective functions: (i) Partition Function (free energy of the ensemble), (ii) Ensemble Diversity and (iii) Nucleotides Composition. The Multiobjective Metaheuristic To Design RNA Sequences (m2dRNAs) proposed in this paper is compared against other RNA inverse folding methods published in the literature, such as RNAinverse, RNA-SSD, INFO-RNA, MODENA, NUPACK, fRNAkenstein, DSS-Opt, RNAiFOLD, antaRNA, ERD, and Eterna players. After a comprehensive comparative study on two well-known benchmarks (Rfam and Eterna100), we conclude that m2dRNAs is capable of obtaining very promising results in terms of both quality of RNA designs and required runtime. The source code of m2dRNAs is available at http://arco.unex.es/arl/m2dRNAs-sourcecode.zip.
- Inverse folding
- Multiobjective optimization