I recently downloaded all the molecular structures (about 13,000) stored in Pherobase, an open library of insect pheromones. I would now like to create a derived, more extensive combinatorial library of compounds that would likely occur in nature, but right now don't happen to be included in the library (e.g., similar aliphatic compounds but with a different chain length or different positions of methyl branches etc).
I know I can create these semi-manually by specifying certain series using SmiLib, or using ChemAxon's Reactor. I was wondering though if there are any good cheminformatic methods to recognize biosynthetic series, which could, e.g., subsume lists of compounds in the library in Markush notation, which could then later be expanded. Or what could be other good ways to expand the library to include other biologically plausible related compounds?
Could Chemaxon's Metabolizer perhaps help, maybe after specifying plausible biosynthetic/enzymatic reactions? Any other ideas?
Just looking in Pubchem for structurally related compounds I think wouldn't work as that would result in too many biologically implausible compounds ... (The reason I want to do this is in the context of making a large QSPR library with in-silico predicted retention indices, and later perhaps also predicted EI-spectra).