Figure 3.

Examples of dynamic programming recurrences. The top part of each example shows relationships in the table and the bottom part the score differences. (left) $\mathrm{\Delta }{\mathsf{\text{G}}}_{\mathsf{\text{nn}}}^{\circ }$ optimization, a nearest-neighbor approximation to the change in free energy of annealing, summing adjacent nucleotide pairs. We must keep track of the third nucleotides of the codons for position i - 1 in order to compute the $\mathrm{\Delta }{\mathsf{\text{G}}}_{\mathsf{\text{nn}}}^{\circ }$ contribution. (right) Run optimization cases: common codon; common codon after substitution; continue and end a run; continue and end one run and start a new one; end one run and start a new one. The run score contribution is computed based on the pattern of nucleotide matches.