Automatically Inferring Atom to Atom Mappings of Chemical Reactions

Uffe Thorsen
University of Southern Denmark

Daniel Merkle
University of Southern Denmark

Christoph Flamm
University of Vienna, Austria

Peter F. Stadler
University of Leipzig, Germany



Content: By modeling molecules as graphs it is possible to infer atom to atom mappings of chemical reactions by solving the minimum edge edit distance problem. This can among other things be used to generate graph grammar rules for use in generative chemistry. Another application is to complete chemical or biological networks by automatically inferring hypotheses for missing chemical reactions missing in the networks. Chemically inspired methods for solving the minimum edge edit distance, both a constructive and a declarative, will be introduced. Included is a discussion on the challenges of and techniques for distinguishing mathematically optimal atom to atom mappings from chemically valid ones.

Back to all abstracts