Open alexbowe opened 10 years ago
This would involve extracting all kmers, then running the packer without adding reverse complements. Could be prototyped easily with a Python tool, but eventually might want to make the reverse complements thing a dynamic flag.
Worth testing without BCALM, since it should be faster (all in memory).
Closing this, since bcalm output is ascii and can easily be converted to a fasta file.
Reopening as latest BCALM may have different format, + we don't accept any ascii input anyway (+ need flag to disable adding revcomps)
Additionally, using STXXL + sorting + set differences, may be possible to implement simple BCALM-like implementation where the minimizers are of k-1 (node) length (which would be worth comparing to BCALM).
BCALM outputs the simple paths, which means only one of each kmer (not its reverse complement). This should make our graph smaller.