News and Announcements » – Pick representative set of sequences


After picking OTUs, you can then pick a representative set of sequences. For each OTU, you will end up with one sequence that can be used in subsequent analyses.

Usage: [options]

Input Arguments:



-i, --input_file
Path to input otu mapping file [REQUIRED]


-f, --fasta_file
Path to input fasta file [REQUIRED if not picking against a reference set; default: None]
-m, --rep_set_picking_method
Method for picking representative sets. Valid choices are random, longest, most_abundant, first [default: first (first chooses cluster seed when picking otus with uclust)]
-o, --result_fp
Path to store result file [default: <input_sequences_filepath>_rep_set.fasta]
-l, --log_fp
Path to store log file [default: No log file created.]
-s, --sort_by
Sort by otu or seq_id [default: otu]
-r, --reference_seqs_fp
Collection of preferred representative sequences [default: None]


The output from is a single FASTA file containing one sequence per OTU. The FASTA header lines will be the OTU identifier (from here on used as the unique sequence identifier) followed by a space, followed by the sequence identifier originally associated with the representative sequence. The name of the output FASTA file will be <input_sequences_filepath>_rep_set.fasta by default, or can be specified via the “-o” parameter.

Simple example: picking a representative set for de novo-picked OTUs:

The script takes as input an ‘OTU map’ (via the “-i” parameter) which maps OTU identifiers to sequence identifiers. Typically, this will be the output file provided by Additionally, a FASTA file is required, via “-f”, which contains all of the sequences whose identifiers are listed in the OTU map. -i seqs_otus.txt -f seqs.fna -o rep_set1.fna

Picking OTUs with “preferred representative” sequences:

Under some circumstances you may have a fasta file of “preferred representative” sequences. An example of this is if you were to pick OTUs against a reference collection with uclust_ref. In this case you may want your representative sequences to be the sequences from the reference collection, rather than the sequences from your sequencing run. To achieve this, you can pass the original reference collection via -r. If you additionally allowed for new clusters (i.e., sequences which don’t match a reference sequence are used as seeds for new OTUs) you’ll also need to pass the original sequence collection to pick a representative sequence from the sequencing run in that case. -i seqs_otus.txt -f seqs.fna -r refseqs.fasta -o rep_set2.fna

Site index

Previous topic – OTU picking

Next topic – plot rank-abundance curve