FAQ: Why do I have more OTUs than ZOTUs?
If you have more OTUs than ZOTUs, this is most likely because more low-abundance unique sequences are clustered by cluster_otus than unoise3. The default minimum size is 2 for cluster_otus and 8 for unoise3. In typical studies, a unique sequence with abundance 8 over all samples will be a singleton or absent in most samples, so the additional OTUs are unlikely to be biologically significant.
To further investigate, you can try cluster_otus using the same -minsize value as unoise3 (the default is 8). The number of OTUs should then be less than or equal to the number of ZOTUs, though it is possible there could be a small excess due to differences in the algorithms such as chimera detection. You can also try clustering the ZOTUs at 97% identity using cluster_fast. This should give a similar number of OTUs to cluster_otus if the same -minsize option was used.
Another approach is to identify OTUs that do not contain any ZOTUs, and ZOTUs that are not found in an OTU. You can do this using usearch_global with a large value of maxrejects to minimize false negatives due to the search heuristics, as shown below.
usearch -usearch_global otus.fa -db zotus.fa -strand plus -id 1.0 -notmatched extra_otus.fa \
-maxrejects 128
usearch -usearch_global zotus.fa -db otus.fa -strand plus -id 0.97 -notmatched extra_zotus.fa \
-maxrejects 128
You can now investigate the extra OTUs and ZOTUs to check why they are missing. First, find the corresponding unique sequence, then use the -uparseout file for cluster_otus or the -tabbedout file for unoise3 to see what happened to that unique sequenece. For example, it might be classified as chimeric. This command finds the unique sequences:
usearch -search_exact uniques.fa -db extra_otus.fa -matched extra_otus_uniques.fa \
-uc extra_uniques.uc
