The starting point for downstream OTU analysis is an
OTU table giving the number of reads per sample per
OTU. This is all you need to perform many standard analyses including alpha and beta diversity.
To get an OTU table in
QIIME classic format, use the -otutabout option
of usearch_global. Some QIIME scripts
require BIOM format, in which case you can use the -biomout option of the
otutab command.
The biom utility, which is included in the
QIIME virtual box, can be used to
convert from QIIME classic to BIOM format. For example,
biom convert -i otutable.txt -o otutable.json --table-type
"OTU table" --to-json
biom convert -i otutable.txt -o otutable.hdf5 --table-type "OTU table" --to-hdf5
You may also need a tree for the OTUs, e.g. for UniFrac analsysis. To make a tree, you can use the cluster_agg command in usearch. For example:
usearch -cluster_agg otus.fa -treeout otus.tree
This generates a tree in Newick format which is compatible with QIIME, mothur and many other programs. Some programs are very picky about labels in Newick files which may cause problems. To get the best compatibility with other software, you should update the sequence labels as follows. Delete any usearch-style annotations (things like size=123;). You can do this with the fastx_strip_annots command. Make sure you use only letters and numbers, don't use anything else (even periods or underscores). Don't use labels longer than 10 characters. Labels like Otu123 are compatible with most programs, though there may still be problems if the integers have too many digits (Otu + 8 digits makes the label >10 characters).
Alpha diversity
You can use the
alpha_diversity.py
script. The -m option specifies the metric(s) to use and the -o option
specifies an output file name.
alpha_diversity.py -i otutab.txt -m shannon -o shannon.txt
alpha_diversity.py -i otutab.json
-m chao1 -o shannon.txt
Beta diversity
You can use the
beta_diversity.py
script. The -m option specifies the metric(s) to use and the -o option
specifies an output directory name.
beta_diversity.py -i otutab.txt -m euclidean -o euclidean