Multi-objective optimization for RNA design with multiple target secondary structures

BMC Bioinformatics

Table 2 The results of two-target designs

l is the nucleotide length of the target set. δ e ₁ and δ e ₂ columns show the energy difference between the ground state and the lowest energy target structure and the difference between the ground state and the highest energy target structure, respectively. The values of δ e ₁ and δ e ₂ are those for the designed sequence with the lowest δ e ₂ in the 100 designed sequences, where the RNA sequence with the lowest δ e ₁ was selected if multiple RNA sequences have the lowest value of δ e ₂. The n ₁ is the number of the designed sequences such that one of the two target structures has the lowest free energy. The n ₂ is the number of the designed sequences in which both target structures have the lowest free energy. The target structures of SV11 are taken from the dataset of RNAdesign [26]. We use the RNA names consistent with the dataset of RNAtabupath (Table 1 of [34]). The target structures of the seventeen sets (from alpha operon to thiM leader) are taken from the dataset of RNAtabupath [34] which contains the dataset of paRNAss [49]; the metastable structures of rb1, rb2, rb3, rb4 [50], and rb5 [51] are taken from the literature. The means and medians are those for the seventeen sets of target structures (the result of SV11 is not used to calculate the means and medians)

ISSN: 1471-2105