Phylum-rank evolutionary distributions of Biosynthetic Gene Cluster classes
MGnify BGCs platform
These results derive from data generated by the MGnift BGCs platform. Full technical detail is in the MGnify BGCs documentation.
The biosynthetic classes
TWe group BGCs by the chemestry of the molecule they encode, following the classes used by the MIBiG reference repository. Some clusters combine chemistries; we label these as hybrids. (Class definitions follow Hanson, J.R.(2003), Natural Products: The Secondary Metabolites (Vol. 17), Royal Society of Chemistry; and Kautsar, S.A. et al. (2020) MIBiG 2.0, Nucleic Acids Research, 48(D1), D454-D458.C)
Image: Terpenes used for chiral pool synthesis. (T. Gaich, J. Mulzer 2012)
Composed of isoprene (C5) units derived from isopentenyl pyrophosphate
Nitrogen-containing compounds derived from amino acids (e.g., ornithine, lysine, tyrosine, tryptophan)
Image: (Wang, S., Lin, S., Fang, Q. et al. A ribosomally synthesised and post-translationally modified peptide containing a β-enamino acid and a macrocyclic motif. Nat Commun 13, 5044 (2022).)
Ribosomally synthesised and Post-translationally modified Peptide
Image: The structure of cyclosporine. (D.K. Daley, K.J. Brown, S. Badal, Chapter 20 – Fungal Metabolites, Editor(s): Simone Badal, Rupika Delgoda, Pharmacognosy, Academic Press, 2017, Pages 413-421, ISBN 9780128021040, https://doi.org/10.1016/B978-0-12-802104-0.00020-2. )
Nonribosomal Peptide
Image: Carbohydrates (Girija S. Singh, Chapter Thirteen – Carbohydrate-based antibiotics: Opportunities and challenges, Editor(s): Vinod Kumar Tiwari, Carbohydrates in Drug Discovery and Development, Elsevier, 2020, Pages 523-559, ISBN 9780128166758, https://doi.org/10.1016/B978-0-12-816675-8.00013-0.)
Carbohydrate-based natural products (e.g., aminoglycoside antibiotics)
Image: cyclitols (J.P. Kamerling, 1.01 – Basics Concepts and Nomenclature Recommendations in Carbohydrate Chemistry, Editor(s): Hans Kamerling, Comprehensive Glycoscience, Elsevier, 2007, Pages 1-38, ISBN 9780444519672, https://doi.org/10.1016/B978-044451967-2/00001-5.)
Catch-all class for clusters encoding metabolites outside main classes (e.g. cyclitols, indolocarbazoles, and phosphonates)
Image: Diterpene and triterpene alkaloids. (Brocksom, Timothy & Oliveira, Kleber & Desiderá, André. (2017). The Chemistry of the Sesquiterpene Alkaloids. Journal of the Brazilian Chemical Society. 10.21577/0103-5053.20170049. )
Clusters combining two or more classes; Polyketide + NRP is flagged separately
The datasets collection
- MIBiG 4.0 – experimentally characterised, published BGCs; the reference set of clusters.
- BacDive – marine bacterial genomes from type strains.
- MGnify Marine Genome Catalogues – genomes assembled from marine water-column and sediment metagenomes.
- MGnify Assembles (v5) – assembled contigs from marine-related metagenomes.
BGC prediction and classification
Three detectors (antiSMASH, GECCO, and SanntiS) annotable each dataset. We merge overlapping predictions on the same genomic interval into one integrated BGC. We derive the integrated BGC class by consensus across detectors. Each detector’s class prediction is collapsed to the shared vocabulary above and the calls are pooled. A single agreed class wins outright. A Polyketide + NRP combination is flagged as that specific hybrid. Any other multi-class combination becomes a general hybrid. Clusters with no usable signal default to Other.
Tree Calculation
Trees were calculated using the CommonTree tool of the NCBI taxonomy datebase. Synonyms were resolved as directed by the CommonTree tool, and other unresolved taxa removed from the tree (this included many BGCs resolving to Candidatus phyla in Bacteria).
