How to process a filtered QIIME fasta

The data

Imagine your collaborator has given you a file lots_o_poo.fna, which contains lines like

>806rcbc0_0 M02171:6:000000000-A6FUV:1:1101:21715:4030 1:N:0:1 orig_bc=GGAGACAAGGGA new_bc=GGAGACAAGGGA bc_diffs=0 TACGGAGGGTCCGAGCGTTAATCGGAATTACTGGGCGTAAAGCGTACGTAGGTGGTTTGTTAAGTTGGATGTGAAAGCCCAGGGCTCAACCTTGGAACTGCATT

This is a quality-filtered QIIME fasta file. The first line identifies the sample name (806rcbc0), the ID for this read in that sample (the 0 after the _), and a whole bunch of other information. The second line is the actual sequence.

Get your ticket to ride

Before boarding the Smile Train, you'll need to create a file q.fst that has usearch-style labels (>sample=sample_name;X where X means that this is the X-th read in sample sample_name). For this example, it would look like >sample=806rcbc0;0. The next read from this sample will have an ID line >sample=806rcbc0;1.

There is a script SmileTrain/util/qiime_to_st_labels.py that can do this for you! Make sure you are not on the head node (maybe qsub -I) and then run

/path/to/SmileTrain/util/qiime_to_st_labels.py /path/to/lots_o_poo.fna > /path/to/q.fst

Take a look to see that q.fst looks the way you expect.

Et voilà.