Specifies a filename prefix for saving clusters. Supported by cluster_fast only. The sequences in each cluster are written to a separate FASTA file. The first sequence in the file is always the centroid. The filename is constructed by appending an integer cluster number 0, 1, 2... to the string specified by -clusters.
The value of -clusters may include a path, which must exist. In the example below, the cluster filenames will be cluster_dir/c_0, cluster_dir/c_1 etc.
usearch -cluster_fast seqs.fa -id 0.90 -clusters