Release Ampliseq Version 3.1.0 · nf-core/ampliseq

nf-core/ampliseq version 3.1.0 - 2026-10-28

SINTAX now accepts the same taxonomic reference databases as DADA2
Consolidation of taxonomic classification can be now done across DADA2, SINTAX and VSEARCH-LCA
Fixed MIDORI-CO1 taxonomic reference for SINTAX & VSEARCH_LCA and removed it for DADA2
DADA2 addSpecies with SILVA 144 and Savont now need fewer computational resources
The pipeline now also works with Nextflow 26.10 (before 26.04 only)

#1123 - --sintax_ref_taxonomy now accepts all databases available for DADA2 (SILVA, GTDB, SBDI-GTDB, RDP, Greengenes2, PR2, GloSED, MIDORI2-CO1, Zehr-nifH), in addition to COIDB and UNITE; requested in #618 (by @erikrikarddaniel)
#1133 - --consolidate_taxonomies most-specific now also picks a winner per ASV across classifiers when DADA2, SINTAX and VSEARCH-LCA are run together; the result is published in consolidated_taxonomy/, reported in the summary report and available as consolidated_across_classifiers_* R objects; consolidated tables have a new classifier column next to database (#1061) (by @erikrikarddaniel, reviewed by @d4straub)

#1122 - README lists what the metro map does not show and the optional steps missing from the pipeline summary; removed the citation and format script of the BOLD Plantae databases, which were removed in 2.15.0 (by @erikrikarddaniel)
#1124 - The --consolidate_taxonomies help and usage documentation now state that DADA2 bootstrap values are not known to be comparable between databases, so score may favour one database for reasons unrelated to accuracy (#1113) (by @erikrikarddaniel)
#1134 - CI runs the conda tests on runners with 32 GB or more of memory: -profile test_savont with GTDB needs about 11 GB for DADA2 assignTaxonomy and was cancelled on 16 GB runners (by @erikrikarddaniel)
#1136 - Launch the AWS tests with seqeralabs/action-seqera-launch and the v2 syntax parser, and stop uploading the debug log as an artifact (by @erikrikarddaniel)

#1119 - DADA2 addSpecies no longer runs out of memory with the SILVA 144 database: SILVA 144 is classified in chunks of 2000 ASVs instead of 10000. A database can set its own chunk size with chunksize in conf/ref_databases.config; --dada_assign_chunksize overrides it (by @erikrikarddaniel)
#1126 - Made params config_profile_* optional/nullable with nextflow v26.10.x (reported by @ebete, fix by @d4straub)
#1132 - The MIDORI2 databases (midori2-co1) for --sintax_ref_taxonomy and --vsearch_lca_ref_taxonomy could not be downloaded because reference-midori.info no longer serves them; they are now downloaded from Zenodo, and the CO1 file is extracted from the archive. midori2-co1 is now release GB273, midori2-co1=gb270 is kept and midori2-co1=gb273 added. midori2-co1 and midori2-co1=gb250 are removed from --dada_ref_taxonomy: GB250 is not on Zenodo, and DADA2 assignTaxonomy cannot use the newer files because they hold more distinct taxonomy strings than its limit of 65,535. Fixes #1130 (by @erikrikarddaniel)

software	previously	now
Savont	0.7.0	0.8.0