The secrets of our bodies—our genes, proteins and other tiny biomarkers scattered throughout cells—have, until recently, exceeded the grasp of our technology to tease out, measure, and make sense of.  But times are changing. After decades of effort and hype, scientists believe they are on the cusp of being able to tell from this data just how healthy each of us is right now, and to augur with some accuracy whether, and when, we might later become ill.

Possibly no scientist has fought harder to reach this goal than physician and biologist Leroy Hood. At the nonprofit Phenome Health, his latest effort to shift the modern paradigm from sick care to well care, he is planning to recruit one million people and collect from them a blizzard of molecular data. Mashed together with medical records, measurements from wearables and other devices, assessments of brain health, and more, Phenome plans to use AI-armed supercomputers to crunch all this data and produce what Hood is calling a “phenomics” profile—a report card designed to untangle and describe a person’s state of health in detail at any given time and arm people with information to stay healthier for longer.

He talked to David Ewing Duncan of Scientific American Custom Media about his new project.

DAVID EWING DUNCAN: Can you describe your idea of how people move from wellness to sickness?

LEROY HOOD: I’m interested in being able to assess and optimize the health trajectory of each individual for body and brain health. This trajectory has three different phases. Most of us start in the wellness phase. Then we go through a transition from wellness to disease. The aim of Phenome Health is to gather powerful statistical evidence for early transitions for every single chronic disease before it’s ever clinically diagnosed. Alzheimer’s is the one we’re starting with. 

Describe the Beyond the Human Genome Project—you’re collecting thousands of data points on each person?

We’re starting by sequencing a million people’s complete genomes, and for each person we’ll also analyze 3,000 proteins, 2,500 metabolites, and all the lipids, plus environmental toxins like mercury. We’re going to do the gut microbiome twice a year, and we’ll use wearable data for things like sleep and heart rate variability. We’ll also assess each person’s brain health—that is, 25 different cognitive features. And much, much more.

You want to raise $10 billion to pay for this, mostly from the U.S. government. Why so much? And is this feasible?

My argument is that for $10 billion we can do a demonstration project that will create the infrastructure so anybody can do this in the future. That’s a database and how to manage it. It’s a biobank and how to manage it. It is modular, computational platforms that handle every aspect of gathering samples, generating data, putting that in the cloud, doing the analysis. We know this will not be easy to raise, but we have supporters in Congress.

How will you make sense of all this data?

There is a new kind of AI called HyperScale AI that’s going to move way beyond statistical correlations and be able to give us directly causal inferences and mechanistic data in a way we could never get before. I think we should take this enormous calculation engine and educate it in several important ways. For instance, let’s feed it everything on PubMed so it has access to all of the published data.

Why do you need to start from scratch with one million more people to test? Wouldn’t it be cheaper to go to, say, the UK Biobank and the 500,000 people they’ve already genetically sequenced?

What’s missing from all the genomes in the UK Biobank are the phenomes. The genome is a kind of source code, and it tells you some interesting things. But you integrate it with phenomes, and it moves you into a completely new dimension.

What about diversity? Most genomes sequenced so far, and much of the data and analysis, are about white people.

We’ve designed our program to reflect the population diversity of the U.S. We’ve done this by having a partner, Guardian Research Network, that interfaces with 120 hospitals and 30 million patients in 13 states across the South and Southwest. They lie across major African American and Latino populations, and others.

How do you plan to take this huge research project and implement it in the clinic and with regular people?

First, you get a small cadre of physicians you’ve convinced, and you work with them in a demonstration project that’s compelling, and it explodes outward from there. There also is cost. We’re going to save enormous amounts of money on drugs. We have biomarkers that will tell you for some drugs which 10 percent of the population they work on, and you can save the other 90 percent the cost of the drugs. We’re confident that the future of health care is well care, and that it will save money and allow people to live longer, healthier lives.

David Ewing Duncan is a journalist who writes for Vanity Fair, Wired, MIT Technology Review, the New York Times, the Atlantic and other publications. He is the author of ten books, most recently, Talking to Robots: Tales from Our Human-Robot Futures (Dutton).

Find out more about Phenome Health’s efforts to transform the future of health care here. Learn more about phenomics, the new science of wellness, in other stories in this special report.