Large-scale metabolic interaction network of the mouse and human gut microbiota

The literature-curated network of microbial interactions within the mouse and human gut portrays a roadmap for mechanistic analyses of microbial community data in biomedical research fields.
Published in Research Data
Large-scale metabolic interaction network of the mouse and human gut microbiota
Like

Do you know that your large intestine is full of microbial cells, regularly excreted in the form of feces? Microbes populate almost every place on Earth, ranging from deep oceans to soils and animal bodies. Among these habitats, our human intestine is the site of a complex and dynamic microbial ecosystem, called the human gut microbiota or microbiome. Recent technologies have revealed that the human gut microbiota may play an important role in our health, while implicated in disorders/diseases such as obesity, diabetes, cancer, irritable bowel syndrome, inflammatory bowel disease, and even autism. I initially learned the concept of the gut microbiota when I was a postdoctoral researcher, and thought that this was an excellent model system for the study of ecological dynamics from a theoretical viewpoint (my original background is theoretical physics). Later, I became more fascinated by the possibility that the gut microbiota research would be practically helpful for improving our real life, by hinting how to modify the gut microbial compositions for achieving better health.

In 2017, I and my team members, especially Dr. Jaeyun Sung (currently at Mayo Clinic), published a global metabolic interaction network of the human gut microbiota. Like a jungle, or a society in some sense, gut microbial cells closely interact with each other and with human tissues. Dietary compounds and host-derived molecules, such as plant fibers and mucins, are broken down by microbes and then released to other microbes as smaller molecules for uptake. Additionally, microbes compete for the same nutrients with each other, or thrive by consuming metabolic byproducts secreted by other microbes. The microbial metabolic byproducts are also absorbed by the intestine, exerting beneficial or harmful effects on the human body. Our 2017 paper presented such a metabolic interaction network of >500 microbial species and 3 human cell types, primarily based on experimental evidence collected from about 400 publications. This network aimed to provide a large overview of microbe-microbe and microbe-host chemical crosstalk inside a human gut, and to help the scientific understanding of microbial innerworkings in particular conditions of interest, as exemplified through type 2 diabetes patient analyses in the 2017 paper.

However, it was not difficult to realize that our network for the human gut microbiota had an obvious limitation in its applicability. Beyond observing simple correlations between microbial compositions and human diseases, more valuable microbiota data with mechanistic investigations of genuine causes of human diseases often come from well-controlled animal experiments, rather than from direct human experiments. In this line, I received collaboration offers for the analysis of data from animal gut microbiota, too. Expanding our network to cover both human and animal gut microbial species would certainly benefit the mechanistic interpretation of the animal microbiota data from various biomedical experiments, and in turn, might even hint how to better design such animal-based experiments. As the most commonly-used animals in biomedical research fields, I decided to choose mice, for our microbiota network update.

Our recent work published in Scientific Data (image produced by Jonathan W. Johnson and Thomas L. P. Martin)

Our team, including Dr. Roktaek Lim, Dr. Jill Cabatbat, and other members, succeeded in expanding the network, which now encompasses metabolic interactions among >800 microbial species and 6 host cell types in the mouse and human gut, as published in our recent paper. In this work, not only was the network expanded to the mouse gut system, but the human gut microbiota network in the 2017 paper was also actively revised to higher quality by cleaning erroneous metabolic interactions and adding new interactions. As a whole, our updated network is a compilation of primarily experimental evidence from about 800 publications—the largest ever, literature-curated network of the mammalian gut microbiota, to my knowledge. Personally, after the publication of the previous network, I moved from Korea to Hong Kong, and have continued this network research.

The current network has a few unique features. Compared to the previous network with only bacterial and archaeal microbes, the current network also includes eukaryotic microbes such as fungi, aiming to cover all three domains of life of microbial cells. In addition, the current network does not only provide the information of known metabolic interactions among microbes and host cells, but also provides the information of which metabolic interactions are not likely to happen, based on the literature. The latter type of negative information would be particularly useful for researchers who may try to infer previously-unknown metabolic interactions by computational algorithms but avoid the retrieval of false-positive interactions not supported by the literature.

Given the huge size of our network (about 8,000 positive and about 900 negative associations between organisms and chemicals), I would like to emphasize the importance of the quality check of individual links in the network by different authors, during this type of the network construction. In our case, we initially re-examined about 30% of the total links to evaluate their quality, and finally, completed the re-examination of the entire links in the network by different authors, followed by the identified error corrections. This huge work was done, partly in collaboration with Dr. Cheol-Min Ghim’s group at UNIST. Based on this collaboration, we further validated our network using the published metabolite levels in the mouse gut and portal vein plasma, as detailed in the paper.

Our network can be freely downloaded from the Dryad Digital Repository, as well as from our website, SymBio.info. In addition, the data package provides the Cytoscape Session (cys) file for visualizing our network (this cys file was produced by Mr. Thomas Martin in my team, and can be opened by free software, Cytoscape). I expect this large-scale network to serve as a global template network which would be adaptable for each specific context of the gut microbiota and thus possibly applicable for developing personalized medicine. For reporting any existing errors in the network or for suggesting any ideas for further network update and expansion, I welcome your feedback through my contact information in SymBio.info.

Please sign in or register for FREE

If you are a registered user on Research Communities by Springer Nature, please sign in

Subscribe to the Topic

Research Data
Research Communities > Community > Research Data

Related Collections

With collections, you can get published faster and increase your visibility.

Remote sensing data for changes in land use

This Collection comprises a series of articles presenting data on changes to land use in urban areas, farmland, forests, and natural environments, as determined using remote sensing techniques.

Publishing Model: Open Access

Deadline: Jan 31, 2024

Ecological data for tracking biological diversity and environmental change

This collection presents data contributions addressing topics in biodiversity and ecology.

Publishing Model: Open Access

Deadline: Jan 31, 2024