The Weizmann Institute of Science's Human Phenotype Project enables the creation of a "digital twin" for each person based on medical data, with the aim of predicting future diseases, identifying medical conditions early, and tailoring personalized treatment - using advanced artificial intelligence.
Before an important meeting or a dramatic decision, we often run simulations in our imagination, and only then choose how to act. When it comes to our health, for example, treating an illness or even choosing a diet, it is much more difficult to predict how each option will affect our body and suit us personally. Using artificial intelligence, a research team from Prof. Eran Segal At the Weizmann Institute of Science, a personal AI "double" allows you to identify the risk of disease, start preventive treatment at an early stage, and even run simulations that will predict which treatment will be most effective. The new development, whichRevealed in an article in the scientific journal Nature Medicine, made possible by The Human Phenotype Project In which the research team, together with partners from around the world, collected in-depth and extensive medical information from more than 13,000 people.
In search of the answer to the question "What makes each of us who we are?", the Human Genome Project was launched in 1990. Until then, only a few genes had been identified, and the project led to the discovery of tens of thousands of genes that shape our traits and found the source of thousands of diseases. However, it is now clear that genes paint only a partial picture. Many of the traits that define us and the diseases that threaten us are related to the environment, the composition of the population of bacteria that inhabit our bodies (the "microbiome"), aging, and more. To complete the picture, in 2018, Prof. Segal from the Department of Computer Science and Applied Mathematics at the Institute initiated the "Human Phenotype Project," which follows thousands of participants who undergo a comprehensive medical assessment every two years for 25 years. The tests review 17 different body systems and include physical measurements, a nutrition diary, ultrasound, bone density testing, voice recording, a home sleep test, continuous two-week monitoring of sugar levels, genetic sequencing and gene expression testing, testing of protein content in cells and metabolic processes, characterization of the microbiome in the intestine, vagina and mouth, and many other tests.
"While the biological age of men increases relatively linearly, in women we detect a jump in age in the fifth decade of life."
"The project was launched in Israel in 2018 with a target of 10,000 subjects," describes Prof. Segal. "Since then, more than 30,000 people have registered and the goal is to reach 100. We wanted to learn more about ethnic, environmental and cultural diversity, so we established a branch in Japan, and we are currently completing the establishment of a branch in the United Arab Emirates, in collaboration with Prof. Eric Shing from the Mohammed bin Zayed University for Artificial Intelligence. We are also expanding the age range. At the beginning, we recruited subjects from 40 to 70, and now younger and older subjects are also being included. The study gave rise to an advanced database, which is both broad in scope and the most in-depth that currently exists in humans. We realized that we had to share it with the scientific community, and today it is digitally accessible to any research group in the world, while maintaining the privacy of the study participants. We believe that the information collected will change the face of medicine."
What is your biological age?
Medicine today is largely based on performing tests and comparing their results to an average range by age and gender. In reality, for each of us, the basic condition and the process of puberty are quite different. A team of scientists, led by Dr. Lee Reicher and Dr. Smadar Shilo from Prof. Segal's lab, developed an artificial intelligence model that studied the typical variation of 17 body systems throughout life and is able to identify abnormalities. The model is based on a platform created by Pheno.AI, a company for artificial intelligence research in the field of health. "The model gives a score to each system and compares it to what is expected according to the subject's chronological age, gender, and weight-to-height (BMI)," explains Prof. Segal. "Based on the deviation from the expected, the model determines the biological age. The older a system is, the higher its risk of disease. For example, by monitoring the sugar levels of the project participants, we discovered what the normal rate of increase in sugar is in men and women over the years. Our model detects deviations from this rate and is able to identify a pre-diabetic condition in 40% of people who were classified as healthy according to currently accepted tests."
The study of biological age has revealed significant differences between the sexes. "While the biological age of men increases relatively linearly, in women we detect a jump in age in the fifth decade of life," describes Prof. Segal. "Menopause is a defining event in many medical aspects and seems to reset the biological age clock. For example, we found that a decrease in bone density is more related to the time that has passed since the onset of menopause than to chronological age. In addition, the indices make it possible to identify the onset of menopause early and plan hormone therapy accordingly." In fact, the Human Phenotype Project has revealed new ways to early diagnose a variety of medical conditions, including breast cancer, inflammatory bowel disease, and endometriosis. It turned out that these diseases are characterized by a change in the composition of the bacterial population in the patient's microbiome, which constitutes a kind of unique "signature" that can be identified.
Still, the project's greatest promise is to be a leap forward on the path to personalized medicine. The scientists aim to achieve this through a single computer model that will combine the information collected from each participant in the Human Phenotype Project, create a "digital double" of him and examine what medical events he is likely to experience in the future and how to prevent them. The model, which is currently under development, led by research student Guy Lutzker, freely learns each person's medical information, and the scientists train it on small prediction tasks. Each time, they hide a certain piece of information from the database and ask it to predict it based on the existing information. Using this training method, a model of generative artificial intelligence (Generative AI) was obtained, which predicts medical events, and can later assemble a personal "health axis" that predicts the person's condition for years to come.
The research team has already developed a model that learned the information about the blood sugar levels of the project participants and was able to predict not only what each of them will have in the future, but also which of the pre-diabetics is at the highest risk of becoming diabetic in the next two years; this prediction makes it possible to prevent or delay the disease at an early stage. In addition, the researchers are already using the digital twin model to test which dietary changes or drug treatments will be personalized for each subject. In the future, the model is expected to encompass all types of information in the database, predict a wide range of events and save trial and error of many treatment directions before finding the right treatment for each person.
"This progress is made possible first and foremost thanks to the community of participants in the Human Phenotype Project. It is a cohesive community of people who are committed to advancing medicine and continuously monitoring their health. We are developing an app that will bring all the information collected to the palm of the participants' hands and will later present them with a personal 'health axis'," says Prof. Segal. "We live in a time of super-rapid change. The worlds of health and medicine will change dramatically in the coming years and become AI-based. Our project is expected to be a leading source of information and innovation in the world, and all thanks to the participants. I would like to take this opportunity to thank each and every one of them - your extraordinary collaboration is the true growth engine of the medical revolution."
More on this subject on Hayadan.com - Science news:
