Computational Framework for the Identification of Bioprivileged Molecules

Bioprivileged molecules are biology-derived chemical intermediates that can be efficiently converted to a diversity of chemical products including both novel molecules and drop-in replacements. Bridging chemical and biological catalysis by bioprivileged molecules provides a useful and flexible new paradigm for producing biobased chemicals. However, the discovery of bioprivileged molecules has been demonstrated to require extensive experimental effort over a long period of time. In this work, we developed a computational framework for identification of all possible C6HxOy molecules (29252) that can be honed down to a manageable number of candidate bioprivileged molecules based on analysis of structural features, reactive moieties, and reactivity of species, and the evaluation of the reaction network and resulting products based on automated network generation. Required input is the structure data file (SDF) of the starting molecules and the reaction rules. On-the-fly estimation of thermodynamics by a group contribution method is introduced as a screening criterion to identify the feasibility of reactions and pathways. Generated species are dynamically linked to the PubChem database for identification of novel products and evaluation of the known products as attractive candidates. Application of the proposed computational framework in screening 29252 C6 species and identifying a list of 100 C6HxOy bioprivileged molecule candidates is presented. Each of the 100 candidate molecules falls into one of nine broad compound classes and is typically composed of carbon atoms with a different chemical environment and, as a result, distinct reactivity patterns. Sensitivity analysis of the parameters used in the filtering steps leading to the candidate molecules that were identified is discussed, and analysis of favorable structural features, reactive moieties, and functionalities of C6HxOy candidate bioprivileged molecules is performed.