Fig. 2
From: LaRA 2: parallel and vectorized program for sequence–structure alignment of RNA sequences
The five steps of the LaRA 2 workflow. 1. Compute individual structure annotation based on base pair probabilities and create all the combinations of pairs to be computed in parallel. 2. Create an alignment graph that satisfies the problem constraints. 3. Formalize constraints as an integer linear program (ILP). 4. The upper boundary for the optimal solution is computed as an ILP that is solved with Lagrangian Relaxation. The lower boundary is obtained with Maximum Weighted Matching. 5. The pairwise sequence–structure alignments are combined with a multiple sequence aligner tool like T-Coffee or MAFFT