Very simple restructuring - if partial extension turns out do be a good idea, then the entire hakmer-ng code can be made MUCH more simple (and probably more efficient, too)... #40
The hakmer-ng algorithem then would simply be:
1.) Extract seeds (allowing up to a small number of mismatches), trimming overlaps in favor of the higher-multiplicity seed
2.) Build superseeds using our simple greedy criterion (optional)
3.) Extend all (super-)seeds by up to a given number of bases, performing partial extension
4.) When printing such an extended (super-)block to a file, compute its MSA
The hakmer-ng algorithem then would simply be: 1.) Extract seeds (allowing up to a small number of mismatches), trimming overlaps in favor of the higher-multiplicity seed 2.) Build superseeds using our simple greedy criterion (optional) 3.) Extend all (super-)seeds by up to a given number of bases, performing partial extension 4.) When printing such an extended (super-)block to a file, compute its MSA