Email
Password
Remember meForgot password?
    Log in with Twitter

article imageEssential Science: Machine learning predicts metabolism

By Tim Sandle     Sep 10, 2018 in Science
London - Researchers are testing out machine learning algorithms which can make a prediction about yeast metabolism based on an assessment of the yeast protein content. This insight will aid scientists to personalize treatments for metabolic disorder patients.
As well as medicine, the research should assist with brewers as well, in terms of giving a basis for brewers to control over the flavor of various beers. The new research comes from The Francis Crick Institute, which is based in London, U.K.
Outside of the Francis Crick Institute  a sculpture of DNA.  A staggering 14 meters high and made of...
Outside of the Francis Crick Institute, a sculpture of DNA. A staggering 14 meters high and made of weathered steel, the sculpture entitled 'Paradigm'.
The focus of the research is upon metabolic processes. Metabolism is the process through which organisms transform nutrients into energy. The process also produces essential molecules. These occur through a progression of chemical reactions.
The recent studies have been conducted using yeast. As yeast metabolizes sugar aerobically, the fungus undergoes a process akin to fermentation to produce alcohol, acids and gases. The process, well established in the food industry, leads to the production of flavor-enhancing compounds which are used in the making of bread; as well as the metabolites that give wine and beer their characteristic tastes.
Saccharomyces cerevisiae is a species of yeast. It is perhaps the most useful yeast  having been ins...
Saccharomyces cerevisiae is a species of yeast. It is perhaps the most useful yeast, having been instrumental to winemaking, baking, and brewing fro a 1000 years.
Douglas Smith
The metabolites produced by a yeast cell are very similar to the metabolites produced by people. While these may have a medicinal use, the way the biochemical process with yeast takes place has not been well understood. This has changed with the new research.
The insight has been achieved through new technology, particularly the application of machine learning algorithms. The researchers used the artificial intelligence to assess the metabolism of the common yeast used in brewing - Saccharomyces cerevisiae.
Saccharomyces cerevisiae is a species of yeast, most likely to have originally been isolated from the skin of grapes. The yeast is one of the most studied eukaryotic model organisms in molecular and cell biology, reflecting the fungus being the most common type of yeast used in fermentation.
Representative image of bacteria
Representative image of bacteria
Geek1
The analysis of the yeast has demonstrated that its metabolism is predictable and there are considerable levels of protein expression information available. For the research, the technologists quantified enzyme expression for 97 different strains of S. cerevisiae. Each strain was known to show differences in metabolism. From this the researchers linked this data to changes in the metabolite concentrations measured.
Machine learning algorithms were then used to detect the complex relationships between changes in gene expression and metabolites produced. This analysis showed the metabolism was controlled by many enzymes acting in concert.
According to lead researcher Dr. Aleksej Zelezniak: “Thanks to machine learning, we now have a better understanding of what controls metabolism, which is good news for brewers looking to create the perfect pint, or for Biotechnologists that use yeast to produce vaccines and other proteins that are medically important.”
The next phase of the research is to take the findings and transfer these to wider research designed to help patients with metabolic diseases.
The research has been published in the journal Cell Systems. The research comes under the title “Machine Learning Predicts the Yeast Metabolome from the Quantitative Proteome of Kinase Knockouts.”
Essential Science
Earthquake damage Santo Domingo/Cumming  Santiago
Earthquake damage Santo Domingo/Cumming, Santiago
Conycampos
This article is part of Digital Journal's regular Essential Science columns. Each week Tim Sandle explores a topical and important scientific issue. Last week we considered a development with seismology. This is a new machine learning approach that can help to predict where aftershocks, following an earthquake, are likely to occur. Aftershocks can often be of the same severity as the original earthquake.
The week before we looked at a new step in terms of developing a universal influenza vaccine.
More about Artificial intelligence, drug development, Metabolism, machine learning
More news from