When I was first approached by our CTO with an idea to write an article about food as medicine, for the share of a second my reaction was “Are you stark raving mad?” However, what I can be confident in is that he would never abandon the scientific approach to seemingly everything.
So is there a scientific approach to eat health? Or, rather, what does Data Science say about food?
Amid today’s obsession with mindfulness and body practices, it’s hard not to notice numerous books, documentaries, and podcasts that explore the association between diet and chronic disease or illness, both somatic and mental.
The present-day application of the famous Hippocrates’ phrase about letting food be your medicine advises us to make better decisions about what to eat in order to prevent disease, maintain health and function better.
Image Credit: pixabay.com
It sounds fair, and a plethora of principles follow, ranging from Ayurveda to genetic programming.
Science probably cannot measure the prana content in your breakfast, but let’s think of more measurable aspects of your diet.
Modern digital personalizes nutritions platforms stake on analyzing anthropometric measurements, blood biomarkers, Nutrigenomics and gut health testing, and here data science can really provide valuable insights.
Anthropometric measurements are used to assess the size, shape and composition of the human body. Height and weight are used for tracking health conditions, drug dosages, as well as nutritional diseases and energy expenditure. As it is impossible to measure every person, their measurements have to be estimated approximately. In this case, different Machine Learning models, starting from simple linear regressions and going to support vector regression, Gaussian process, and artificial neural networks are used to predict height and weight more accurately.
Biomarker testing is the core of personalized medicine. In its broadest sense, the term “biomarker” refers to any of the body’s molecules that can be measured to assess health, including molecules from blood, body fluids, and tissue. Biomarker testing looks for molecular signs of health so that doctors can plan the best care. Applying machine learning techniques is critical to assess the importance of different biomarkers and to collect the best marker combinations that have the biggest impact on the person’s response to treatment or — in case of nutrition — to personalize the diet according to the metabolism.
Image Credit: unsplash.com
Nutrigenomics refers to the use of biochemistry, physiology, nutrition, genomics, proteomics, metabolomics, transcriptomics, and epigenomics to seek and explain the reciprocal interactions between genes and nutrients. These interactions are believed to aid the prescription of personalized diets according to one’s genotype mitigating the symptoms of existing diseases or preventing future illnesses, especially in the area of Nontransmissible Chronic Diseases. Data Science is playing a major role in nutrigenomics analysis, as it helps to find patterns predicting future outputs, to study the present status of nutraceuticals in the market, and to control the customers’ responses to them. Using cluster detection, memory-based reasoning, genetic algorithms, link analysis, decision trees, and neural net, huge amounts of data can be handled in a short period of time.
Gut health tests help to understand what nutrients and toxins are being produced by gut microbiome. Gut contain bacteria, viruses, fungi, phages, yeast, parasites, etc. that may be active to a varying degree and produce nutrients or toxins from the digested food influencing our mood, digestion, immune system, skin, fitness, and many more. Artificial intelligence helps process all information about the gut microbiome to accurately assess risks of development of diseases and to recommend the exact foods you should be eating and foods you should be minimizing with the goal to keep your gut in balance.
Intelligence (AI) and Machine Learning (ML) techniques to increasingly applied to identify patterns and correlations in between the gathered user data and specific health-related conditions. These correlations, already known at some forums as Digital Biomarkers, will empower preventive health approaches, including personalize diet development. In this sense, we have enabled AI with our world’s most complex data set: the data set of biology. It is true that what we have learned about genomics on the one hand and nutrients on the other is incredible, but what we don’t know still outweighs what we have learned. Biology is extremely complex, and we are only on our way to understanding the impact of the food we eat on our health — but now at least we can make decisions based on proven data.
The science behind our diet may be sophisticated, but the basic principles are simple: understand your body, its microbiome and genes, and use food rather than supplements that suit your organism the best. And the main principle of any health lifestyle: “Everything in moderation, including moderation.”
