The digital tool, known as AGORA, uses data to create separate computer models that mimic the metabolism of 773 bacterial strains found in the human gastrointestinal tract.
When used with other datasets, these simulations can further explain how metabolic interactions in the gut microbiome are influenced by diet.
The AGORA models may also be of enormous value in addressing host–microbe co-metabolism questions that personalised nutrition or medicine research consistently raise.
"With our models, we can search, in a targeted manner, for metabolic pathways that are fundamentally important to the microbiome in the gut, and we can work out what could trigger diseases when these metabolic processes go wrong," said co-author Dr Ronan Fleming, who lead the Systems Biochemistry group at the LCSB.
"The AGORA models will now allow us to study the impact of host-microbiome interactions in specific diseases or to use them in the emerging field of personalised medicine."
Gut bugs and disease
The disruption of gut bacterial functions, such as amino acid and vitamin production, breakdown of indigestible plant polysaccharides and production of metabolites have been linked to chronic diseases including type 2 diabetes and obesity.
Studies into these microbiome communities has led to enormous data sets being generated that require systems biology tools to mine and provide additional insight.
Existing tools are limited as they cannot identify the contribution of each bacterial species to the metabolic functions that characterise the whole gut microbiome.
The team from the University of Luxembourg demonstrated AGORA’s predictive capability by using two models, Bacteroides caccae (B. caccae) ATCC 34185 (a fibre-degrading organism common in microbiomes of Western individuals) and Lactobacillus rhamnosus GG (LGG), a common human probiotic strain.
They predicted that B. caccae would grow on a specific culture medium (DMEM 6429) supplemented with vitamin K, hemin, and arabinogalactan in the absence of oxygen. They also predicted that the medium would not support growth of LGG.
The AGORA models also predicted that B. caccae would supply LGG with alanine, asparagine, and nicotinic acid, while LGG would provide lactate to B. caccae.
Using metabolomic analysis, the team confirmed the secretion of numerous metabolites by the two strains grown individually, including alanine secretion by B. caccae thus supporting the predicted cross-feeding.
“We want to understand how the microbes modulate human metabolism when we modify our diet,” explained lead study author Dr Ines Thiele, head of the Molecular Systems Physiology team at the Luxembourg Centre for Systems Biomedicine (LCSB) based in the University of Luxembourg.
“This may give us clues as to how we may prevent, or even treat, diseases, for example by identifying dietary supplements that could modify the interactions within a diseased gut microbiome to imitate the metabolic functions of a healthy one."
Responding to the paper's findings, the International Probiotics Association (IPA) commented on the benefits of reconstructing the intestinal microbiome's genomic and metabolomic capacity using sequences of 773 members of the intestinal microbiota.
"The advantage of this approach is that one can predict the contribution of individual members to the overall intestinal metabolism. This is benefit to the probiotic industry; albeit in the long term.
"The approach allows for the understanding of the interaction between endogenous microbes and probiotics and may provide insight into the mechanisms of probiotic action and may even help predict interactions in silico. Though this does probably not lead to prediction of health effects.
"Furthermore, the technique may help with the selection and cultivation of the so-called next generation probiotics.”
The collection of predictive metabolic models are now available to researchers.
Source: Nature Biotechnology
Published online ahead of print, DOI: 10.1038/nbt.3703
“Generation of genome-scale metabolic reconstructions for 773 members of the human gut microbiota.”
Authors: Ines Thiele et al.