Our computational biology section develops robust models to untangle the intricate interactions between cellular machinery, metabolic pathways, and experimental data, ultimately making wet-lab work more insightful and predictable. By integrating multi-omics datasets, including transcriptomics and metabolomics, within mathematical frameworks, we generate predictive simulations for cell-free systems. These in silico tools help us tackle key challenges, like understanding how nuclease activity competes with transcription and translation or how the accumulation of toxic metabolites can limit protein yields. Our aim is to build a streamlined, “single-button” modeling platform that enables researchers to forecast system behavior, speed up biosensor design, and guide rational strain engineering with confidence.
Example Projects
Assessing Structural Uncertainty of Biochemical Regulatory Networks in Metabolic Pathways
This project aims to tackle a critical challenge in systems biology: understanding how metabolites regulate complex biochemical pathways, especially when experimental data is noisy or incomplete. Using computational models, we assess how different data qualities affect our ability to identify these regulatory networks accurately. By applying biochemical systems theory and analyzing common network structures, we strive to pinpoint the key factors that influence the success of these models. This work not only deepens our understanding of metabolic regulation but also helps develop more reliable tools for predicting pathway behavior, ultimately aiding the design of better experiments and advancing biotechnological applications.
SCOUR (Stepwise Classification of Unknown Regulation)
Understanding how metabolites regulate complex biochemical networks is a major challenge in systems biology, particularly due to the noisy and incomplete nature of experimental data. Our team developed SCOUR, a stepwise machine learning framework that systematically identifies regulatory interactions in metabolic systems using metabolomics and fluxomics datasets. By synthetically generating the training data needed to identify reaction flux regulators, SCOUR streamlines the process of testing and validating potential regulatory interactions. This approach reduces the time and effort required for experimental validation, helping to advance our ability to model and engineer metabolic pathways in both existing and novel organisms.
Read More
- Assessing structural uncertainty of biochemical regulatory networks in metabolic pathways under varying data quality
- Towards inferring absolute concentrations from relative abundance in time-course GC-MS metabolomics data
- Diverse classes of constraints enable broader applicability of a linear programming-based dynamic metabolic modeling framework
- SCOUR: a stepwise machine learning framework for predicting metabolite-dependent regulatory interactions