$ micca otu --help
usage: micca otu [-h] -i FILE [-o DIR] [-r FILE]
                 [-m {denovo_greedy,closed_ref,open_ref}] [-d ID] [-n MINCOV]
                 [-t THREADS] [-c] [-g {dgc,agc}] [-s MINSIZE]
                 [-a {both,plus}]

micca otu assigns similar sequences (marker genes such as 16S rRNA and
the fungal ITS region) to operational taxonomic units (OTUs).
Trimming the sequences to a fixed position before clustering is
*strongly recommended* when they cover partial amplicons or if quality
deteriorates towards the end (common when you have long amplicons and
single-end sequencing).

micca otu provides the following protocols:

* de novo greedy clustering (denovo_greedy): sequences are clustered
  without relying on an external reference database, using an
  approach similar to the UPARSE pipeline (doi: 10.1038/nmeth.2604):

  1. predict sequence abundances of each sequence by dereplication,
     order by abundance and discard sequences with abundance value
     smaller than --minsize parameter;

  2. greedy clustering. Sequences are considered in order of
     decreasing abundance. Distance (DGC) and abundance-based (AGC)
     strategies are supported, see doi: 10.1186/s40168-015-0081-x
     and doi: 10.7287/peerj.preprints.1466v1 (parameter '--greedy').
     Therefore, the candidate representative sequences are obtained;

  3. (optional) remove chimeric sequences from the representatives
     (parameter '--rmchim') performing a de novo chimera detection;

  4. map sequences to the representatives.

* closed-reference clustering (closed_ref): sequences are clustered
  against an external reference database and reads that could not be
  matched are discarded.

* open-reference clustering (open_ref): sequences are clustered
  against an external reference database and reads that could not be
  matched are clustered with the 'denovo_greedy' protocol.


* otutable.txt: OTU x sample, TAB-separated OTU table file,
  containing the number of times an OTU is found in each sample.

* otus.fasta: FASTA file containing the representative sequences (OTUs);

* otuids.txt: OTU ids to original sequence ids (tab-delimited text file)

* hits.txt: three-columns, TAB-separated file:

  1. matching sequence
  2. representative (seed)
  3. identity (if available), defined as:

              matching columns
     -------------------------------- ;
     alignment length - terminal gaps

* otuschim.fasta (only for 'denovo_greedy' and 'open_ref' when --rmchim
  is specified): FASTA file containing the chimeric otus;

optional arguments:
  -h, --help            show this help message and exit

  -i FILE, --input FILE
                        input fasta file (required).
  -o DIR, --output DIR  output directory (default .).
  -r FILE, --ref FILE   reference sequences in fasta format, required for
                        'closed_ref' and 'open_ref' clustering methods.
  -m {denovo_greedy,closed_ref,open_ref}, --method {denovo_greedy,closed_ref,open_ref}
                        clustering method (default denovo_greedy)
  -d ID, --id ID        sequence identity threshold (0.0 to 1.0, default
  -n MINCOV, --mincov MINCOV
                        reject sequence if the fraction of alignment to the
                        reference sequence is lower than MINCOV. This
                        parameter prevents low-coverage alignments at the end
                        of the sequences (for 'closed_ref' and 'open_ref'
                        clustering methods, default 0.75).
  -t THREADS, --threads THREADS
                        number of threads to use (1 to 256, default 1).
  -c, --rmchim          remove chimeric sequences (for 'denovo_greedy' and
                        'open_ref' OTU picking methods).
  -g {dgc,agc}, --greedy {dgc,agc}
                        greedy clustering strategy, distance (DGC) or
                        abundance-based (AGC) (for 'denovo_greedy' and
                        'open_ref' clustering methods) (default dgc).
  -s MINSIZE, --minsize MINSIZE
                        discard sequences with an abundance value smaller than
                        MINSIZE after dereplication (>=1, default 2).
                        Recommended value is 2, i.e. discard singletons (for
                        'denovo_greedy' and 'open_ref' clustering methods).
  -a {both,plus}, --strand {both,plus}
                        search both strands or the plus strand only (for
                        'closed_ref' and 'open_ref' clustering methods,
                        default both).


De novo clustering with a 97% similarity threshold and remove
chimeric OTUs:

    micca otu -i input.fasta --method denovo_greedy --id 0.97 -c

Open-reference OTU picking protocol with a 97% similarity
threshold, without removing chimeras in the de novo protocol step
and using 8 threads:

    micca otu -i input.fasta --method open_ref --threads 8 --id 0.97 \
    --ref greengenes_2013_05/rep_set/97_otus.fasta