Collective Data, Individual Health

BIOTECHNOLOGY AND THE HUMAN GENOME PROJECT

Hood played a major role in developing the DNA sequencer used in the Human Genome Project, a publicly funded international effort to study all of the DNA of a single human genome sequence, and in the 1990s, was at the forefront of the project. Hood has continued his work into an expansive future by employing big data approaches for studying the origins of health and advancing more targeted disease prevention and treatment — precision human health.

Hood began his lecture by describing how the field of phenomics has been made possible due to major advancements over the past 20 years.

“There were no good tools for generating enormous amounts of data that could be used in deconvoluting everything,” Hood said.

Today, with big data on the rise, truly individualized approaches to health — focusing on wellness rather than upon disease treatment and management — will soon be possible through phenomics, Hood said.

Phenotypes are more specific than genotypes in that all of an organism’s observable characteristics are influenced by its genotype or specific DNA sequence, its environment and its behavior. Each individual has a unique phenotype. When studying the phenotypes of a million or more people, scientists can provide precise recommendations for disease interventions such as diet, exercise and even medication.

Hood described the changes over the past 50 years as a series of paradigm shifts, beginning with the merger of biology and engineering in the 1970s and then movement to the Human Genome Project, which was launched in 1990 and completed in 2003.

The Human Genome Project cost around $3 billion to produce a single human genome sequence. This major scientific advance has led to breakthroughs in the study of immunology and genetics. However, it was just the beginning of a shift toward more individualized health made possible through electronic health records and systems biology approaches to health.

PHENOMICS AND PERSONALIZED HEALTH

Through phenomics, big data is employed to predict how people with specific phenotypes will respond to disease interventions and how their health recommendations can be custom designed.

“How we find the biological age is enormously complicated,” Hood said. “We look at a series of algorithms. Then, we move on to data analysis, which is the most exciting part. Moving from correlation to causal inference makes a much bigger difference.”

Hood envisions that aging itself will be considered more relative in terms of a person’s biological age versus their chronological age. The major features of behavioral interventions may include exercise, diet, stress management and intermittent fasting recommendations.

Body mass index (BMI), as one example, misclassifies up to 30 percent of people, but it is currently the best tool that is widely available, Hood said. With better, more intuitive systems and health metrics such as data-driven BMI, physicians would gather much more comprehensive information, and individuals could obtain insights about how behavioral changes like starting a new strength training routine or adjusting their sleep schedules will impact their overall health. He describes this concept as “P4 health”: predictive, preventive, personalized and participatory.

Hood identified several major problems facing the United States health care system, which he said lags among high-income countries due to a lack of wellness-focused medicine. Those problems include quality of care, an aging population that will require more resources, value-based care and diversity, equity and inclusion.

Here inclusion refers to including a broad enough sample size to account for variability according to ethnicity and also including groups that may not have had the access or ability to participate in clinical trials in the past. In other words, scientists need data that is representative of the entire human population to help the largest number of people. Where the Human Genome Project was concerned with a single genome, Phenome Health’s Human Phenome Initiative will collect data from more than a million individuals that span the diversity represented in the U.S. population.

Researchers will integrate newly collected data and contextualize known mechanisms captured in Google and UC San Francisco knowledge graphs. Through Google’s Biomedical Data Commons, there are a staggering 50 billion nodes, 850 billion edges, 215,000 variables and 566 properties.

Next, through hyperscale artificial intelligence, an educated system similar to ChatGPT will be fed individual data to derive actionable recommendations that will then be delivered to clinicians. Ultimately, each person can have a digital twin to simulate the effectiveness of certain disease interventions — without needing to implement these disease interventions on the individual.

At the event, Adams noted that school leaders from across the University including Dr. Inga Musselman, provost, vice president for academic affairs and the Cecil H. Green Distinguished Chair of Academic Leadership, and chemistry professor, and Dr. Steven Small, dean of the School of Behavioral and Brain Sciences and Aage and Margareta Møller Distinguished Professor in Behavioral and Brain Sciences, were present.

Adams and Dr. James Yurkovich, who currently serves as chief innovation officer at Phenome Health as well as being a member of the Industrial Advisory Council for the University’s Department of Bioengineering and part-time systems engineering lecturer in the Jonsson School, introduced Hood. The scale and scope of an endeavor such as what Hood presented are massive.

“The numbers and statistics Hood shared echo why we must have diversity in science,” said Adams, also a professor of systems engineering. “For our students – keep dreaming, keep pushing, get out of your comfort zone. The work we need to do is transdisciplinary.”

Discover More