Meta-Alignment with Crumble and Prune: Partitioning very large alignment problems for performance and parallelization

Table 5 Prune results for twelve alignment problems from the Rfam database.

¹ Pecan was run with default parameters.
² FSA was run with the --exonerate, --anchored, --softmasked, and --fast flags.
³ MUSCLE was run with default parameters.
⁴ MAFFT was run with --treein option.
⁵ SATé was run with the -t option but limited to two iterations. We found that more iterations did almost nothing for accuracy.
^aThe majority of problems were unable to be aligned due to running out of memory.
The run-time and agreement score of Prune alignments of twelve RNA alignment problems from the Rfam database. The average time and agreement over all twelve problems are shown. Pecan, FSA, and MUSCLE were used as the underlying alignment method of Prune. MAFFT and SATé were also tested to provide comparison. We were unable to apply Pecan without using Prune because of memory issues. Using Prune, we were able to use Pecan to solve these alignment problems. Prune achieved a very large speedup with little loss of accuracy. Other alignment methods achieved a large speedup but more accuracy was lost.