Fig. 1

a 1926 MIBiG BGCs with corresponding structures were queried to antiSMASH, NPClassifier, and ClassyFire to retrieve BGC and chemical compound classifications. The arimetamycin MIBiG entry (BGC0000199) is given as an example with part of its BGC and the arimetamycin B chemical structure, for which the chemical compound classes are determined and the genome-based and structure-based classifications are matched. Please note that the NPClassifier and ClassyFire ontologies have more class terms than shown here. b Sankey diagram showing the relationships between matched MIBiG and NPClassifier pathway class terms. For reference, 722 BGCs had a Polyketide MIBiG class, 548 of which matched to NPClassifier pathway Polyketides. c Example scoring matrix between MIBiG classes and NPClassifier pathways, where each count is divided by the total occurrence of the class, in this case the column, to compute a score (see Eqs. 1 and 2 in the ‘Implementation’ section)