Radhakrishnan Mahadevan

Professor

New Projects Available

• Machine learning for bio catalyst engineering
• Bioengineering Mine Tailings for Nickel Extraction
• Hybrid Deterministic and Machine Learning for Metabolic Engineering

Currently accepting graduate students? Yes

• PhD & MASc

Memberships

American Institute of Chemical Engineering
American Chemical Society
American Society for Microbiology
Biophysical Society

Research Interests

Research Interests: systems biology, synthetic biology, machine learning, metabolic engineering, protein engineering, gut microbiome, microbial communities. Applications include bioproducts including renewable chemicals and fuels, CO2 conversion to fuels and chemicals to fight climate change, and bioprocess optimization. Medical applications include multi-scale whole body metabolic models for nutrition, engineering probiotics for therapeutics to treat inflammatory diseases such as IBD.

Metabolic engineering using systems and synthetic biology

Availability of genome-scale metabolic models can accelerate the optimization of metabolism for the synthesis of biochemicals and fuels. A key requirement for sustainable development is the ability to synthesis chemicals and fuels from renewable feedstocks. We are using these genome-scale models to optimize microbial metabolism for the production of fuels and chemicals. In addition, we are also developing novel synthetic biology tools using genetic circuits such as toggle switch for the dynamic control of metabolic pathways in order to optimize metabolism effectively for biochemicals production. In addition, we have set up robotic platforms to accelerate synthetic biology and automate cell engineering. We are collaborating with several companies to advance this area of biochemicals and fuels.

Systems analysis and engineering of biological processes

Recent advances in experimental and computational technologies have enabled the detailed characterization of biological systems. In particular, the molecular components of these systems including the list of genes, proteins they encode, and compounds that interact with these proteins can be determined. This availability of tools to analyze system-wide changes at the level of the genes, proteins, and metabolites has created significant opportunities to understand cellular functions, and to ultimately design processes in a systematic way for applications in industrial and medical biotechnology (e.g. metabolic engineering, bioprocess optimization and control). The research interests of our group involve the development and utilization of dynamic mathematical models of biological systems for improved design, optimization and control.

Genome-scale models of cellular processes and microbial communities

Although detailed models of cellular processes have been constructed in the past, research in this area has attained a new dimension in the last few years due to the development of novel high-throughput experimental techniques for both sensing and manipulating cellular processes at a molecular level. As an example, both steady state genome scale models and smaller dynamic models of metabolism of several industrially important organisms including Escherichia coli have been developed in the past. More recently, such models have been developed for a metal reducing bacteria (Geobacter sulfurreducens) with applications in bioelectricity and bioremediation and have been used to rationally engineer the metabolism for improved electricity generation. However, further research is required to extend such models of metabolism to represent the inherent dynamics of biological systems and to account for the increased complexity in multi-cellular organisms and microbial communities.

Machine learning for protein modeling and enzyme engineering

Advances in machine learning has enabled the accurate prediction of protein structures from sequences using methods such as AlphaFold and RoseTTAFold leading to intriguing opportunity to apply such learning for predicting protein function from sequence and for machine learning guided protein engineering. While existing methods are able to predict structures from sequence reasonably, function prediction is still unresolved. We have developed accurate biosensors for small metabolites and are leveraging these sensors along with in house automation and collaboration with JGI to generate data and are developing novel machine learning methods for protein language modeling using natural language processing algorithms.

Optimization and control of biological processes

Several engineering disciplines (e.g., mechanical, electrical, & chemical) routinely use quantitative models for design and optimization of processes of interest. However, such rational approach to design and optimization has been possible in the life science only recently due to the lack of predictive large-scale models of biological processes in the past. Research activities in our group include the design of dynamic model-driven engineering strategies for biological process optimization and control across different length and time scales (i.e., from microscopic (intracellular) processes to macroscopic (bioreactor) processes). Applications can include metabolic engineering (e.g., increasing the rate of electrical current in microbial fuel cells, designing dynamic gene manipulation strategies for increased product yields), biomedical engineering (drug design and dosage, personalized nutrition and medicine), bioreactor control and optimization (designing optimal substrate and inducer feeding strategies), and bioremediation (determining the spatiotemporal substrate addition strategies to effectively stimulate microbial activity). This project includes collaborations with industry.

Selected Publications

A. Pandit, E. Harrison, R. Mahadevan, “Engineering Escherichia coli for the Utilization of Ethylene Glycol”, Microbial Cell Factories, 2021, 20(1), 1-17.

Y. Liu, J. Chen, D Crisante, JM Lopez, R. Mahadevan, “Dynamic Cell Programming with Quorum Sensing Controlled CRISPRi", ACS Synthetic Biology, 2020, 9(6), 1284-1291.

R. Chaudhury, R. Mahadevan, “Towards Engineering Smart Probiotics”. Current Opinion in Biotechnology, 2020, 64, 199-209.

C. Lieven et al., “MEMOTE for standardized genome-scale metabolic model testing”, Nature Biotechnology, 2020, 38, 272-276.

N. Venayak, A. von Kamp, S. Klamt, R. Mahadevan, “MoVE identifies metabolic valves to switch between phenotypic states”, Nature Communications, 2018, 9 (1), 5332.

K Nemr etal.,Engineering a short, aldolase-based pathway for (R)-1, 3-butanediol production in Escherichia coli, Metabolic engineering, 2018, 48, 13-24.

Pandit AV, Srinivasan S, Mahadevan R. Redesigning metabolism based on orthogonality principles. Nature Communications 2017;8:15188.

J.C. Joo, et al. Alkene hydrogenation activity of enoate reductases for an environmentally benign biosynthesis of adipic acid. Chemical Sciences, 8 (2016), pp. 1406-1413.

Xu, Z., Wu, J., Song, Y.S. & Mahadevan, R. Enzyme Activity Prediction of Sequence Variants on Novel Substrates using Improved Substrate Encodings and Convolutional Pooling. Proceedings of the 16th Machine Learning in Computational Biology meeting, in Proceedings of Machine Learning Research, 2022, 165:78-87