Abstract

Constraint-based Reconstruction and Analysis (COBRA) is a widely used systems biology approach that relies on molecule-resolved genome-scale reconstructions of a target organism’s metabolism to simulate the flow of metabolites in silico. COBRA models can be tailored condition-specific through integration of omics data including transcriptomics, proteomics, metabolomics, and metagenomics. Such condition-specific models have promising applications for the interrogation of changed fluxes in pathways associated with cardiometabolic diseases such as obesity and metabolic syndrome. Here, I will present genome-scale modeling approaches to interrogate the interactions between diet, the microbiome, and human metabolism.

Previously, we have published AGORA2, a resource of human microbe genome-scale reconstructions accounting for 7,302 genome-scale reconstructions of human gut, skin, oral, and vaginal microbes. By mapping relative abundances retrieved from 16S rRNA or shotgun sequencing data onto AGORA2, personalized microbiome models can be constructed. Early-life exposures including the gut microbiome influence metabolic programming and can have long-lasting impact. Birth mode affects the composition of the gut microbiome in early life. To investigate the impact of birth mode through genome-scale modeling, we used metagenomic datasets from 20 infant gut microbiomes at five days, one month, six months, and one year of age and built personalized microbiome models from AGORA2 for each infant and time point. We found that the metabolic output of the gut microbiome differed between vaginally delivered infants and those delivered through Cesarian section at the earliest time points. The metabolic capabilities of the infant gut microbiome were also distinct from those of adult gut microbiomes.

A whole-body model of metabolism that can be contextualized with diet, physiological parameters, and the gut microbiome, Harvey, has also been developed. I present a workflow to use Harvey for the prediction of organ-level effects of obesity on metabolism. I retrieved transcriptomic and methylome data from a cohort of 37 morbidly obese patients. First, I used a liver-specific reconstruction to build and interrogate personalized liver models for patients. I correlated metabolic fluxes with methylome data and clinical parameters. Finally, I contextualized Harvey with the clinical parameters of patients and simulated metabolic fluxes in obesity across organs. Taken together, personalized modeling of metabolism on the organ level and in the microbiome allows us to predict individual-specific metabolic fluxes and analyze them in the context of clinical parameters such as birth mode, weight, or insulin resistance.