Complex Reaction Networks

We have developed methods for automated generation of reaction mechanisms of complex systems that allow kinetic models of substantive detail to be built. Molecules are represented as graphs and matrices, and operations on these representations allow reaction to be carried out, molecule uniqueness to be determined, and properties to be calculated. We have applied our methodology to a wide range of different problems, including production of silicon nanoparticles, biochemical transformations, polymerization and depolymerization, and tropospheric ozone formation. While the chemistries we have studied are seemingly very disparate, applying a common methodology to study them reveals that there are many features of complex reaction networks that are ubiquitous.

Discovery of Novel Biochemical Transformations

Carbon-based compounds are marked for a transition. Chemicals derived from biomass are viewed as a companion, or even as a successor, to compounds derived from petroleum. Growing energy consumption will place heavy pressure on petroleum resources, and it is generally accepted that relying on finite petroleum reserves is not a satisfactory policy for the long term. How do we begin to move in the direction of biochemical conversion of renewable resources? There are numerous possible chemicals as targets, yet their production using bioprocessing may not have been demonstrated yet. Even if a chemical is currently produced using biochemical reactions, there are opportunities for process improvement via the use of alternative routes. To navigate among the vast number of possible combinations of chemicals and processes for producing them, computational thinking is key. However, it is not sufficient to simply search databases of known compounds and reactions. The diversity of chemistry and biochemistry suggests that there are pathways that have not been discovered yet, and it is even possible that novel compounds have yet to be synthesized. We are developing a cyber-enabled exploration platform that will discover and analyze complex biochemical reaction networks that:

• Present a novel but superior route to a chemical that is already produced biochemically
• Offer a novel route to produce a specialty chemical biochemically that is typically produced via traditional organic synthesis
• Propose a biochemical synthesis route to a novel compound

Automated network generation that defines and implements the chemistry of what we have coined "generalized enzyme functions" based on knowledge compiled in existing biochemical databases is employed. Graph theory is used to represent compounds and chemical actions of enzymes mathematically so that product compounds can be created from reactant compounds through simple operations. The output is a set of compounds and the pathways connecting them. We are also developing methods and linking them with the pathway generation algorithms that allow the compounds and pathways to be screened for their novelty and the pathways to be evaluated for their potential for implementation. Our method for automated generation of pathways creates novel compounds and pathways that have not been reported in biochemical or chemical databases. Thus, our method goes beyond a survey of existing compounds and reactions and provides a radical departure from the conventional approaches practiced to develop novel biochemical processes.


The generation of novel pathways to known and novel fuels and chemicals relies on representing the breaking and formation of chemical bonds as mathematical operators. Molecules are examined to see if they have the necessary functional groups to undergo the reaction encoded by the operators. The functional group requirements can be made less strict, allowing the operators to be more general, thereby resulting in a greater degree of novelty.


Hatzimanikatis, V., Li, C., Ionita, J.A. and Broadbelt, L.J., "Metabolic Networks: Enzyme Function and Metabolite Structure", Current Opinion in Structural Biology, 2004, 14, 300-306.

Li, C., Henry, C.S., Jankowski, M.D., Ionita, J.A., Hatzimanikatis, V. and Broadbelt, L.J., "Computational Discovery of Biochemical Routes to Specialty Chemicals", Chem. Eng. Sci., 2004, 59(22-23), 5051-5060.

Hatzimanikatis, V., Li, C., Ionita, J.A., Henry, C.S., Jankowski, M.D. and Broadbelt, L.J., "Exploring the Diversity of Complex Metabolic Networks", Bioinformatics, 2005, 21(8), 1603-1609.

González-Lergier, J., Broadbelt, L.J. and Hatzimanikatis, V., "Theoretical Considerations and Computational Analysis of the Complexity in Polyketide Synthesis Pathways", J. Am. Chem. Soc., 2005, 127(27), 9930-9938.

Henry, C.S., Jankowski, M.D., Broadbelt, L.J. and Hatzimanikatis, V., "Genome-Scale Thermodynamic Analysis of Escherichia coli Metabolism", Biophysical J., 2006, 90(4), 1453-1461.

González-Lergier, J., Broadbelt, L.J. and Hatzimanikatis, V., "Analysis of the Maximum Theoretical Yield for the Synthesis of Erythromycin Precursors in Escherichia Coli", Biotechnology and Bioengineering, 2006, 95(4), 638-644.

Henry, C.S., Broadbelt, L.J. and Hatzimanikatis, V., "Thermodynamics-based Metabolic Flux Analysis", Biophysical J., 2007, 92(5), 1792-1805.

Feist, A.M., Henry, C.S., Reed, J.L., Krummenacker, M., Joyce, A.R., Karp, P.D., Broadbelt, L.J., Hatzimanikatis, V. and Palsson, B.O, "A genome-scale metabolic reconstruction for Escherichia coli K-12 MG1655 that accounts for 1261 ORFs and thermodynamic information", Nature Molecular Systems Biology, 2007, 3, Art. No. 121.

Jankowski, M.D., Henry, C.S., Broadbelt, L.J. and Hatzimanikatis, V., "Group Contribution Method for Thermodynamic Analysis of Complex Metabolic Networks", Biophysical Journal, 2008, 95, 1487-1499.

Finley, S.D., Broadbelt, L.J. and Hatzimanikatis, V., "Thermodynamic Analysis of Biodegradation Pathways", Biotechnology and Bioengineering, 2009, 103: 532-541.

Finley, S.D., Broadbelt, L.J. and Hatzimanikatis, V., "Computational Framework for Predictive Biodegradation", Biotechnology and Bioengineering, 2009, 104: 1086-1097.

Mechanistic Modeling of Oxidation Chemistry

A multi-scale modeling approach was used to describe oxidation chemistry in the liquid phase. A library of kinetic correlations was first established that is suitable for estimating rate coefficients and activation energies for condensed-phase free-radical oxidation in hydrocarbons with the specific target application of modeling the thermal degradation of lubricating oils. Structure-reactivity relationships for 17 different reaction families relevant to lubricant degradation are reported. Nine structure-reactivity relationships have not been reported elsewhere in the literature. All of the structure-reactivity relationships are determined by using reaction rate coefficients and activation energies available from experiment or calculated via quantum chemistry and transition state theory. The high-level CBS-QB3 method was used in order to calculate activation energies and heats of reaction for several hydrogen transfer reaction sub-families. It was shown that due to the variety of radicals and substrates present in propagation reactions in lubricant degradation, unique reaction sub-families are important for accurately capturing the different reactivity trends observed. Then, automated mechanism generation was used to study the condensed-phase oxidation of decane and octane as a means to gain insights into the degradation of lubricating oils. Specific reaction rules are proposed that enable automated mechanism generation to be used, for the first time, to study condensed-phase free radical oxidation of large substrates. Models of decane oxidation were generated, and very good agreement with available experimental data was achieved. The optimized parameters were then used to generate the first predictive models of octane autoxidation.


Multiscale modeling of hydrocarbon oxidation links the atomic scale with the process scale. Quantum chemical calculations are carried out to unravel mechanisms and predict rate coefficients. These are combined with techniques for automated mechanism generation to create detailed kinetic models that can be linked with reactor design equations to predict behavior at the macroscopic scale.


Pfaendtner, J., Yu, X. and Broadbelt, L.J., "Quantum Chemical Investigation of Low-Temperature Intramolecular Hydrogen Transfer of Hydrocarbons", J. Phys. Chem. A, 2006, 110(37), 10863-10871.

Pfaendtner, J. and Broadbelt, L.J., "Elucidation of Structure-Reactivity Relationships in Hindered Phenols Via Quantum Chemistry and Transition State Theory", Chem. Eng. Sci., 2007, 62(18-20), 5232-5239.

Pfaendtner, J., Yu, X. and Broadbelt, L.J., "The 1D Hindered Rotor Approximation", Theoretical Chemistry Accounts, 2007, 118(5-6), 881-898.

Pfaendtner, J. and Broadbelt, L.J., "Mechanistic Modeling of Lubricant Degradation Part 2: The Autoxidation of Decane and Octane", Ind. Eng. Chem. Res., 2008, 47(9), 2897-2904.

Pfaendtner, J. and Broadbelt, L.J., "Mechanistic Modeling of Lubricant Degradation Part 1: Structure-Reactivity Relationships for Free-Radical Oxidation", Ind. Eng. Chem. Res., 2008, 47(9), 2886-2896.