Align.seqs

It looks like that the align.seqs command functions the same way as the aligner of greengeens (http://greengenes.lbl.gov/cgi-bin/nph-NAST_align.cgi).
Could you let me know if there is any difference between results from each aligner?

Yes, it’s very similar. I am finishing responding to the reviewers’ comments, so hopefully the paper will be published describing the changes in PLoS ONE soon. Here’s a snapshot of the differences…

  1. You do not have to use greengenes’s reference alignment. We have produced much better alignments using a homemade SILVA reference alignment (improved accuracy and speed)
  2. You do not have to use blast to do the searching for the best template or to do the alignment. It turns out that using blast actually really sucks for this application - it is slow, trims sequences, doesn’t find what it’s supposed to and produces a bad alignment (improved flexibility, speed, accuracy, etc.)
  3. You can use kmer searching with any sized kmer to do the searching (improved flexibility, speed, and accuracy)
  4. You can use the needleman-wunsch or gotoh pairwise alignments, which are very fast (improved speed)
  5. You can set the match/mismatch and gap opening/extension scores to do the alignment (improved flexibility)
  6. You can use multiple processors on your computer (improved speed)
  7. You don’t have to be online or wait for the greengenes servers (improved flexibility and speed)
  8. It’s open source

Hopefully, you’ll be convinced that this is an improved version. The paper runs through a number of simulations and also shows how well the mothur aligner performs using short sequence reads.

Pat