Scientists at Boston University Schools of Medicine and Public Health have developed machine learning methods that are capable of identifying and characterizing chemicals that disrupt metabolism. Her research is published in Environmental Health Perspectives.
If you sit on your sofa too much, you can gain weight. Sounds like common sense, doesn’t it? However, the weight gain may not be due solely to an excessively sedentary lifestyle. Rather, it could be caused by exposure to certain chemicals that might be in your furniture.
These chemicals are known as Metabolic Disrupting Chemicals (MDCs) or “obesogens” and can be found in various household items and throughout the environment. As the name suggests, MDCs can trigger changes in a person’s metabolic processes and create a predisposition to weight gain by stimulating the formation of fat cells (adipocytes).
Scientific research has only recently begun to examine exactly what types of fat cells are produced – there are different types – as a result of exposure to such chemicals. “This is an important question because not all fat cells are ‘created equal’,” says Dr. Stefano Monti from the Department of Medicine at Boston University. “White fat cells store energy and contribute to obesity. Brown and brite (brown-on-white) fat cells burn energy and reduce obesity. Our previous work suggests that environmental chemicals are more likely to stimulate white fat cell formation.”
Monti explains that there is a correlation between increased production (and exposure to) environmental MDCs and the rapid increases in obesity and metabolic disorders observed in humans. “Recent studies have shown that the increase in BMI in recent years is not simply due to excessive calorie intake and / or insufficient energy expenditure,” he adds.
To limit our exposure to and use of these potentially harmful chemicals, we need to know what and where they are, which has proven difficult. However, Monti is the corresponding author of a new study that successfully used the machine learning approach to successfully identify and characterize MDCs in a number of unclassified chemicals.
What is machine learning?
Machine learning, a branch of artificial intelligence (AI), uses data and algorithms to replicate the way people learn. For example, to learn a task, people repeat it and perform the task until it is optimized. The same thing happens with machine learning; with improved accuracy every time.
Why machine learning?
Why use machine learning in this context? The decision was based on a desire by Monti and his colleagues to develop an approach that is both unbiased and data-driven. Machine learning enabled the team to “learn” effectively from previous research studies. ‘We’ profiled ‘a number of more than 60 chemicals with known effects (that is, known to be either obesogenic or non-obesogenic) and used them to’ train ‘a computer model to predict their metabolic disruptive potential “describes Monti.
In the profiling phase of the experiment, pre-adipocyte cells – obtained from mice – were treated with each of the chemicals and mRNA extracted from them. Next, the mRNA was sequenced using RNA sequencing (RNA-seq) methods for transcriptional analysis. This process provided the researchers with information on how the cells’ genes had responded to the chemical stress. “These RNA sequencing profiles, along with the known chemical labels, were fed into a computer model that was trained to differentiate between the two classes and then applied to the classification of unlabeled chemicals,” says Monti.
The RNA-seq profiles provided information on the effects of short-term exposure to the chemicals, while the markings (e.g., obesogenic or non-obesogenic) were used to provide longer-term exposure effects. Therefore, the machine learning model was trained to use the short-term expression profiles to predict the possible long-term exposure effects of the unlabeled chemicals. Monti emphasizes that this is a subtle but important point.
The design of the experiment builds on previous work, the Carcinogenome Project, which aimed to identify potential carcinogens. “Together, the two studies provide a conceptual, experimental, and computational framework (i.e., a comprehensive ‘recipe’) with general applicability for screening large quantities of chemicals for their potential long-term side effects, including, but not limited to: metabolic disorders and carcinogenicity,” says Monti.
The full effects of MDC exposure
The research group would like to emphasize that the application possibilities of their latest study go beyond the specifics of the method used and its predictive capabilities. Chemicals profiled in the study also included drugs used to treat metabolic disorders. Their methodology therefore enabled scientists to study more closely how these drugs affect a cell’s metabolism. “This understanding, in turn, will be critical to developing more effective and targeted drugs with minimal side effects,” says Monti.
Identifying a chemical as an MDC is just the first step, explains Monti: “We selected two of the high-level predictions (tonalide and quinoxyfen, two commonly used pesticides) and conducted an extensive functional validation that conclusively confirmed their harmful effects on human fat formation However, more testing would be required to determine any [regulatory] Action, ”he concludes.
Stefano Monti spoke to Molly Campbell, Science Writer for Technology Networks.
relation: Kim S, Reed E, Monti S, Schlezinger J. A data-driven transcriptional taxonomy of adipogenic chemicals for the identification of white and British adipogens. Vicinity. Health perspective. 2021. 129 (7): 077006. do: 10.1289 / EHP6886.