Skip to main content

Table 2 Comparison rBAN versus s2m

From: rBAN: retro-biosynthetic analysis of nonribosomal peptides

  rBAN Smiles2Monomers
a) Monomers mapping Based on molecule fragmentation through common monomer linking bonds Based on mapping of monomers and selection of best tiling
b) Light matching Positions of double/triple bonds are ignored Implicit hydrogens and bond order are ignored
c) Heterocycles treatment Accounts for NRP cyclisation patterns initiating oxazoles and thiasoles formation Does not include any rule/pattern for heterocycles
d) Presence of new monomers Unmatched regions left unannotated and potentially identified in discovery mode Matches the most similar monomers in a given database and leaves out uncovered atoms
e) Graph serialization Labelled edges with bond type and directed in accordance to functional groups in each side Unlabelled edges
  1. a) To map the monomers rBAN fragments the molecule and matches the results against the monomer database. S2m computes the combinations of monomers that fit in the molecule. b) To enable tautomer identification during the matching process rBAN omits the positions of the double bonds in the monomer, but it keeps considering those, becoming more restrictive than its analog mode in s2m, in which neither the implicit hydrogens nor the bonds order are taken into account. c) Characteristic NRP structural patterns such as heterocycles are specifically targeted in rBAN but not in s2m. d) When a region cannot be matched because of the absence of the monomer in the database, rBAN leaves the whole region unannotated (with the option of recurring to the discovery mode), while s2m tries to match the most similar monomer even if this is a wrong match and it implies leaving unannotated atoms. e) The monomers graph from rBAN has the edges labeled specifying the type of bond and its direction. s2m does not provide bond labels