ClusterCompare is a new algorithm for cluster comparison, taking into account presence of biosynthetic profiles, NRPS/PKS module counts and layouts, and gene functions along with the sequence identity and synteny used by the clusterblast module.
The range of scores for any pairing is between 0 and 1, with 1 being a theoretical perfect score. When comparing multiple protoclusters against a single region, the score may be higher and mousing over each protocluster's result will show the 0-1 score for that particular protocluster.
ClusterCompare has been initially released with a comparison database for MIBiG 2.0
--cc-mibig), but includes a script for easily generating custom databases
(which can be used with
The precision mode of RREfinder was added and can be run with
--rre. This will
detect and annotate RiPP Recognition Elements within RiPP protoclusters.
Annotations based on the TGIRfam database can now added to a run (with
These annotations will be present in the genbank output and
also be shown in the gene detail panel on the top right of the HTML output.
A method of running antiSMASH analyses on genomic areas that antiSMASH does not
detect a cluster in is now possible via the use of sideloaded annotations. These
annotations, provided in a JSON format, can contain protoclusters and/or subregions
along with extra details to attach to those regions. One or more sideload external
annotation files with
The clusterfinder module has been removed due to being often misinterpreted or misleading.
- a number of additions and changes were made to detected RiPP protoclusters, see the glossary page for details
- JSON output now includes a simple representation of all detected areas
- output filenames can be customised with
--output-basenameinstead of defaulting to the input name
- Pfam domains for the region are shown in a detail tab similar to the existing NRPS/PKS domains
- Pfam domain hits for a single gene are now included in the gene detail panel of the HTML output
- added a button to download SVG images in the HTML output
- detection rules can now create and use aliases for groups of profiles
Fixes and small changes#
- improved error messages for invalid inputs
- improved handling of non-standard genbank inputs
- a workaround has been added for a MacOS specific issue using python 3.8+ with
--cpusset higher than 1
Numerous other small changes and fixes were made internally, for a full list see the git shortlog.