Quantifying the activity of enzymes operating within the large-scale biochemical network is a fundamental challenge in Systems Bio(tech)nology. Here the unknown quantities must be inferred from models that are incomplete and measurements that involve errors. For such challenges, Bayesian analysis using Markov Chain Monte Carlo (MCMC) has become a standard tool.
For addressing high dimensional parameter inference problems with Bayesian statistics, powerful MCMC methods have emerged, for example the MCMC differential evolution and the Riemann Manifold Langevin Monte Carlo method. Because of the specific structure of the inference problems occurring in metabolic models, direct application of these MCMC algorithms is not possible.
In this project, you will bring MCMC methods into the setting of metabolic flux inference and with inspiration from existing algorithms, develop tailored MCMC algorithm(s). The ensuing algorithm(s) will be implemented in an existing C++ framework, validated and benchmarked with a realistic case study.
The focus of the project can develop either more in the mathematical theory of MCMC or practical implementation of code for the Jülich supercomputers.