Set marker_gene: 16S or marker_gene: 18S in the parameters file.
PEMA supports both ASV inference and OTU clustering; set clustering_algo in the parameters file:
swarm: Swarm infers ASVs. Its options are in the swarm section; the usual setting is
differences: 1 with fastidious: true. Higher values of differences merge more sequences into each ASV.dada2: DADA2 infers ASVs from the read pairs, as in its tutorial:
fastp filters the pairs without merging them; DADA2 learns the error rates and denoises the forward and the reverse reads separately,
merges the denoised pairs and removes the chimeras. Its options are in the dada2 section (e.g. pool, min_overlap); to truncate the reads to
fixed lengths, as DADA2 recommends, set trunc_len_forward and trunc_len_reverse there (applied by fastp, only for DADA2). The read counts after each step are in dada2/dada2_read_tracking.tsv.vsearch: VSEARCH clusters OTUs, at the identity threshold of the vsearch section
(e.g. 0.97), after removing chimeras de novo, as in the VSEARCH pipeline.min_oligotons removes ASVs with a low total abundance after Swarm or DADA2. With VSEARCH, it removes unique sequences with a low total
abundance before clustering (as in PEMA v2.x), which removes many more reads: use a low value, e.g. 2, with min_oligotons in the vsearch section (VSEARCH only).
With classifier: crest, PEMA assigns the taxonomy with crest4, an LCA-based classifier,
against one of its databases (database in the crest section):
silvamod138pr2: SILVA 138 with PR2, for 16S and 18S;ssuome: the crest4 SSU database, for 16S and 18S.crest4 downloads the database the first time it is used; set mount_crest_dbs to a folder of yours (e.g. ~/.crest4) to keep it there,
so it is downloaded only once. The search uses BLAST or VSEARCH (search_algo); min_score, score_drop and min_smlrty set how strict the
assignment is (see the crest section of the parameters file).
You can also use your own reference database; see Training the CREST classifier.
The phylogeny-based taxonomy assignment (EPA-ng) and the phyloseq analysis of PEMA v2.x are not part of the Nextflow-based PEMA;
the final ASV/OTU table (tax_assigned_table.tsv) can be loaded in phyloseq, or any other R or Python package, directly.