Genome scale metabolic network modelling for metabolic profile predictions

Area of Science:

• Biochemistry
• Systems Biology
• Computational Biology

Background:

• Metabolic profiling (metabolomics) analyzes small molecules (metabolites) in biological samples for human health studies.
• Unlike genomics, no single metabolomics setup captures the entire metabolome, necessitating complex experimental design.
• Computer-assisted approaches can streamline the design of metabolomics experiments.

Purpose of the Study:

• To develop a computational method for predicting potential biomarkers of metabolic perturbations.
• To accelerate the design and interpretation of metabolomics studies.
• To identify metabolites likely to be differentially abundant in disease states.

Main Methods:

• Utilized a constraint-based modeling approach with flux simulation on genome-scale metabolic networks.
• Implemented SAMBA (SAMpling Biomarker Analysis) to simulate and compare metabolite exchange fluxes between baseline and perturbed conditions.
• Ranked differentially exchanged metabolites as potential biomarkers based on simulated flux distributions.

Main Results:

• Demonstrated a strong correlation between simulated metabolic exchange profiles and experimentally detected differential metabolites in plasma.
• Validated findings using patient data from the OMIM database and metabolic trait-SNP associations from mGWAS studies.
• Successfully identified potential biomarkers indicative of specific metabolic perturbations.

Conclusions:

• The SAMBA approach effectively predicts potential biomarkers for metabolic perturbations using in silico flux simulations.
• This method aids in understanding disease mechanisms and metabolite differential abundances.
• Recommends novel metabolites for further experimental investigation in metabolomics research.