Figure 1From: Genetic algorithm learning as a robust approach to RNA editing site predictionGA optimized weights for six variables in REGAL. We selected six variables and utilized the GA to optimize the weights for these variables. The greater the importance of a variable, the higher the value as shown here. Variables were abbreviated as follows: 1 = codon_transition, transition probability for codon pre- and post-edit; 2 = +1_nucleotide, nucleotide in +1 position relative to candidate C; 3 = hydrophobicity, likelihood that edit will yield a more hydrophobic amino acid than the unedited codon; 4 = amino_acid_transition, transition probability for amino acid pre- and post-edit; 5 = codon_position, position of the candidate edit site with respect to the codon (i.e. first, second or third position); and 6 = -1_nucleotide, nucleotide in the -1 position relative to candidate C.Back to article page