Glossary

Short definitions of the platform’s terms

Each entry links to the page where the term is explained in full. New to the platform? Start with Key Concepts.

Adjacency (domain adjacency)

The neighbouring domain pairs along a cluster’s proteins. Used alongside domain overlap to measure cluster similarity, capturing arrangement rather than just membership. See Scores & Metrics.

Assembly

The genome or metagenome a cluster was found in — an isolate genome, a metagenome-assembled genome (MAG), an environmental assembly, or a single validated BGC region. In the discovery platform, assemblies provide context (organism, source, taxonomy, biome) for the iBGCs they contain. See Key Concepts.

BGC (biosynthetic gene cluster)

A stretch of DNA whose genes work together to produce a natural product. See Key Concepts.

BGC class

A normalised label for the kind of biosynthetic machinery — Polyketide, NRPS, RiPP, Terpene, Saccharide, and hybrids. See BGC Classes.

Biome

The environment an assembly came from, following the GOLD ecosystem classification (e.g. Marine, Marine sediment). See Biome Ontology.

CHAMOIS

The tool that predicts a cluster’s product chemistry (as ChemOnt classes) from its protein sequences. See How the Data Is Built.

ChemOnt

A hierarchical chemical-structure ontology (from ClassyFire) used to label predicted product chemistry. See ChemOnt.

Composite similarity

The weighted combination of domain overlap and adjacency overlap used to compare clusters. The basis for clustering, find-similar, and novelty. See Scores & Metrics.

Detector

A BGC detection tool — antiSMASH, GECCO, or SanntiS. Each produces source predictions that are integrated into iBGCs. See Key Concepts.

Domain (protein domain)

A functional unit within a protein (InterPro/Pfam), e.g. a ketosynthase or adenylation domain. The platform’s currency for comparing clusters. See Protein Panel.

Domain novelty

The fraction of an iBGC’s domains that are unique within its own GCF. “—” for single-member families or clusters with no domains. See Scores & Metrics.

GCF (gene cluster family)

A group of iBGCs with similar domain content — a “natural product family”. Identified by a dotted path (e.g. 42.7.3). See Gene Cluster Families.

iBGC (integrated BGC)

The platform’s central unit: overlapping source predictions on one contig consolidated into a single candidate cluster. See Key Concepts.

MIBiG

The curated database of experimentally validated BGCs. Loaded as validated iBGCs; the reference set novelty is measured against. See MIBiG & Validated BGCs.

Novelty

1 − similarity to the nearest validated cluster. Higher = less like anything known. “—” when no validated reference exists. See Scores & Metrics.

Partial

An iBGC that runs off the edge of its contig and may be incomplete. Placed onto the maps and scored by its nearest complete neighbours. See UMAP.

Reference / Compare

The two iBGC detail panels: the reference is pinned (right-click → set as reference); the compare panel updates on each left-click. See iBGC Detail.

Region (assembly mode)

A processing mode where the input is a known BGC (one per contig); detectors are off. Used for validated sets. See How the Data Is Built.

Shortlist

The browser-held set of iBGCs you collect, turned into a report and exports. Holds up to 1,000 iBGCs. See Shortlists.

Source / collection

The dataset an assembly came from (e.g. a marine catalogue, BacDive, MIBiG). A filter and a roster column. See What’s in the Catalogue.

Sørensen–Dice

The set-overlap coefficient used to score domain and adjacency similarity: 2·|A∩B| / (|A|+|B|). See Scores & Metrics.

Submitted

An iBGC from an assembly you uploaded with Load Asset, shown in amber. See Load Asset.

Type strain

An assembly corresponding to a strain deposited in a public culture collection (DSMZ, ATCC, …) — i.e. one you can obtain and culture. Flagged in the roster and detail panels. See What’s in the Catalogue.

Validated

An iBGC from an experimentally characterised reference (e.g. MIBiG). The yardstick for novelty; novelty 0 by definition. See MIBiG & Validated BGCs.