A new approach that uses artificial intelligence (AI) shows how to use microorganisms in the body and molecules in cells to predict human health outcomes, according to researchers at Penn State College of Medicine and Southwestern Medical Center. the University of Texas. They say it could improve the accuracy of predicting the development of human diseases, such as inflammatory bowel disease and diabetes.
The human microbiome is made up of billions of microorganisms, such as fungi and bacteria that live in the body, usually in the gut, and affect overall health. These organisms, along with the metabolome -; or molecules found within cells and tissues -; they have a significant impact on medical research.
The present study proposes to learn useful features of data sets that measure both the microbiome and the metabolome and use them to substantially improve the accuracy of risk prediction in data sets that only measure the microbiome. The results present a statistical learning and non-invasive AI-based approach that uses the intestinal microbiome that could identify people at high risk for disease.
Until now, due to cost constraints, only a handful of studies measured data from both the microbiome and the metabolome. Most studies only measured microbiome data without including data on metabolomes, which limited its usefulness in predicting disease risks. According to the researchers, combining the microbiome and the metabolome can more accurately predict disease outcomes and lead to a better understanding of the mechanisms of the disease.
Non-invasive approaches based on deep learning have enormous potential to improve the diagnosis and prediction of human disease risk. Combined with high-performance technologies, such as DNA sequencing, it offers a cost-effective approach that identifies patients at risk and rapidly advances precision medicine. ”
Dajiang Liu, senior co-author, professor and vice president of public health science research and biochemistry and molecular biology, and interim director of the AI initiative at Penn State College of Medicine
The scientists proposed a new integrative modeling framework called Microbiome-Based Supervised Controlled Learning Framework (MB-SupCon). In implementing the new method, they studied intestinal microbiome and metabolome data in stool samples from 720 patients to predict factors associated with type 2 diabetes.
According to the researchers, MB-SupCon surpassed the existing machine learning methods and was very accurate in predicting the state of insulin resistance (84%), gender (78%) and race (80%) of patients.
When the researchers used MB-SupCon in a study of inflammatory bowel disease, they observed similar benefits. According to the researchers, this non-invasive and cost-effective method could be widely used to predict health outcomes in several disease studies.
“The human microbiome is an important modifiable risk factor for human disease,” said lead co-author Xiaowei Zhan, a member of Southwestern Medical Center at the University of Texas. “Our approach helps identify bacteria that influence disease risk. Modifying these bacteria may be a valuable new approach to treating human disorders that were previously not easily treatable.”
Researchers Sen Yang, Shidan Wang, Ruichen Rong, Jiwoong Kim, Bo Li, Andrew Y. Koh and Guanghua Xiao of the University of Texas Southwestern Medical Center; Yiqing Wang of Southern Methodist University; and Qiwei Li of the University of Texas at Dallas contributed to this research.
Source:
Penn State College of Medicine
Magazine reference:
Yang, S., et al. (2022) MB-SupCon: Microbiome-based predictive models through supervised contrastive learning. Journal of Molecular Biology. doi.org/10.1016/j.jmb.2022.167693.