What is it about?
Biomarkers are key variables in the research and development of new methods for training prognostic and classification models based on advances in artificial intelligence. Biomarkers contribute significantly towards “personalized medicine”. Specifically, with the deployment of deep neural networks, a link between two very important metabolic routes has been explored. The pathogenic route is investigated through a classification based on the Body Mass Index (BMI), via identifying patterns in the biochemical profile compiled via a simple blood test. The dietary route is looked into in a similar way via linking the diet intake with the biochemical profile elements that seem to be more influential in defining our samples weight class. By separating the weight classes in different groups and by balancing the data, an improvement in accuracy was achieved. Through the deployment of different combinations for data features, the main conclusion was that the quantity of data is the main benefactor in the deployed system predictability. There is extensive bibliography regarding metabolism, dietary profiling, metabolomics and related research that utilizes statistical methods for correlation and causation purposes. Our aim is to design a model that can automate a mundane procedure, to create a fast and easily applicable method for scaled applications with the use of deep neural networks and pattern recognition techniques.
Featured Image
Photo by National Cancer Institute on Unsplash
Why is it important?
The doctrine of the “one size fits all” approach in the field of disease diagnosis and patient management is being replaced by a more per patient approach known as “personalized medicine”. In this spirit, biomarkers are key variables in the research and development of new methods for prognostic and classification model training based on advances in the field of artificial intelligence. Metabolomics refers to the systematic study of the unique chemical fingerprints that cellular processes leave behind. The metabolic profile of a person can provide a snapshot of cell physiology and, by extension, metabolomics provide a direct “functional reading of the physiological state” of an organism. Via employing machine learning methodologies, a general evaluation chart of nutritional biomarkers is formulated and an optimised prediction method for body to mass index is investigated with the aim to discover dietary patterns.
Perspectives
A bigger data set is guaranteed to improve classification ability and accuracy. Interpreting and extracting content from metabolic data sets is extremely demanding and represents an important area of research. A more challenging approach would be to integrate metabolomics with genomics and proteomics. In summary through this research: – A link between blood and weight with the use of simple deep neural networks was established. – The semantic grouping can be a stepping stone for the evolution of the proposed classification procedure. – The more analytical approach of the data offer a clear view of what is tested and through that related bias can be calculated. The identification of patterns in complex data sets can lead to the discovery of new biomarkers and add precision in medical classification and patient management. This and other related research is currently under way and will be reported elsewhere in the near future.
Prof.-Dr. George A. Tsihrintzis
University of Pireaus
The Greek philosopher Pythagoras profoundly said that no one is free who has not obtained the empire of himself. Artificial intelligence is a pathway for us to simplify the observations we have made about how the systems that we are involved in and are consisted of, work and through that find ways to optimise them. This collaborative effort was driven by the need to understand the human body and to create rules related to how our body works, for the machine to translate into functions. The process automation can further facilitate the obtaining of knowledge and with that knowledge improve the understanding of our biology and thus improve health outcomes.
Dimitrios P. Panagoulias
University of Piraeus
Read the Original
This page is a summary of: Nutritional biomarkers and machine learning for personalized nutrition applications and health optimization, Intelligent Decision Technologies, January 2022, IOS Press,
DOI: 10.3233/idt-210233.
You can read the full text:
Contributors
The following have contributed to this page
