3D visualization model can study metabolism! The paper is on the cover of Science Signaling

Remember the diagram of human metabolism in your high school biology textbook?

Although it seems not complicated, its research has been a difficult problem that has troubled scientists for many years. Recently, scientists have built an interactive 2D/3D visualization model to better study metabolism.

On March 24, 2020, local time, Jonathan L. Robinson's team from Chalmers University of Technology in Sweden published their research results entitled An atlas of human metabolism online in the journal Science Signaling. They built the most complete metabolic model in history called "Human 1" and its supporting portal website Metabolic Atlas, which systematically displayed the human metabolic lineage.

It is worth mentioning that this achievement also appeared on the cover of the journal.

Metabolism refers to the exchange of substances and energy between the body and the environment, as well as the self-renewal process of substances and energy in the body, including anabolism (assimilation) and catabolism (dissimilation). Simply put, metabolism is the replacement of old substances with new substances.

In fact, metabolism is a function of cells and a symbol of a person's vitality. If metabolism is abnormal, it may cause many health problems, such as obesity, diabetes, hypertension, heart disease and cancer. Currently, doctors mainly diagnose diseases by screening metabolic markers in patients' blood and urine, and researchers have also begun to conduct research on targeted metabolic processes for disease treatment.

Leifeng.com learned that it is now possible to measure thousands of metabolites simultaneously, but it is not easy to understand the metabolism in human cells as a whole.

The strategy proposed by scientists is to construct genome-scale metabolic models (GEMs), which have actually been used in industrial applications involving Saccharomyces cerevisiae and Escherichia coli. They aim to understand metabolism, design new cellular targets (such as biofuel production) and improve yields. They are our way to fully understand the basis of human metabolism and provide a framework for the comprehensive analysis of omics data.

As early as 2005, scientists began the long scientific research journey of building GEM, and all efforts began with the development of "Recon1" and "Edinburgh Human Metabolic Network" (EHMN).

We can think of them as the "forefathers" of two independent model series, the Recon series (including Recon1, 2 and 3D) and the HMR series (including HMR1 and 2). These two model series have been largely merged with each other during the update period and can be used to treat diseases such as disorders, diabetes, fatty liver and cancer.

However, even with the above two model series, GEM still faces challenges in further development:

• the use of nonstandard identifiers for genes, metabolites, and reactions;

• duplication of model components;

• Errors in early model iterations;

• Updates to the model are opaque and difficult to coordinate.

In fact, these challenges are the starting point for Jonathan L. Robinson's team's research .

As shown in the figure below, the team built a new and arguably the most complete GEM lineage in history, named Human1, which was obtained by integrating and managing the above-mentioned Recon and HMR model series.

Leifeng.com learned that the entire Human1 development process was carried out systematically in the GitHub repository to enable tracking of all updates.

Specifically speaking, this process is a big project - in order to integrate the two model lineages of Recon and HMR, the research team deleted 8185 duplicate reactions and 3215 duplicate metabolites, revised 2016 metabolite formulas, reestablished 3226 reaction equations, corrected the reversibility of 83 reactions, and disabled/deleted 576 inconsistent/unnecessary reactions.

In addition, to demonstrate the versatility and predictive accuracy of Human1, the researchers comprehensively analyzed the transcriptional data of 33 tumors and 53 healthy tissues, conducted gene importance studies involving more than 620 different cell types, and used enzyme-constrained GEM (ecGEM) derived from Human1 to predict nutrient exchange and growth rate of the NCI-60 cell line.

At the same time, Human1 is also able to accurately simulate cell growth and metabolite exchange rates given limited flux information.

Notably, the research team created the Metabolic Atlas, a portal accompanying Human1.

The following figure is a manually drawn two-dimensional map covering 6793 non-transport/non-exchange reactions (90%), 4027 metabolites (97%) and 3316 genes (91%) in the human body. These maps are integrated with transcriptional data from the Human Protein Atlas (HPA), and gene expression levels from 37 different tissue types can be superimposed on the transcriptional data.

The following is a screenshot of the 3D atlas.

3D effect poke link:

https://www.metabolicatlas.org/explore/map-viewer/human1/compartment/golgi_apparatus?dim=3d

Using the website, other research teams can further explore the contents of Human1 and upload their own transcription data to further advance the visualization.

It can be seen that this research result has improved the ability to model metabolic pathways related to human health and disease, providing relevant researchers with the opportunity to further improve metabolic models at the human genome scale, and will become a shared resource for human health and disease research in the future.

References:

https://baike.baidu.com/item/%E6%96%B0%E9%99%88%E4%BB%A3%E8%B0%A2/108770?fr=aladdin

https://stke.sciencemag.org/content/13/624/eaaz1482

https://www.metabolicatlas.org/