The cost of sequencing the human genome — about $1,500 as of 2015 and continuing to decrease — is one millionth of what it was just 14 years ago when the first human genome was sequenced. This plummeting cost curve puts Moore’s Law to shame. We expect other personal measurements to exhibit a similar decline. As costs drop, we will be able to extend very sophisticated testing and personal health evaluation — an approach we call “scientific wellness” — to everyone, leading to a democratization of health care unimaginable even a few years ago. No more will leading-edge technologies necessarily be the most expensive ones. Everyone can benefit from the knowledge hidden in their personal dense dynamic data cloud — based on their genome, their microbiome, their medical history, extensive blood analysis of chemistries, proteins and metabolites, and even the daily input from fitness trackers and home scales. Analyzing and integrating these data will lead to actionable possibilities that will permit individuals to improve their wellness and avoid disease.
Two new strategies have facilitated this revolution in medicine: systems medicine (global or comprehensive approaches to disease) and the data-driven personalized measurements embodied in what I call “P4 medicine”: care that is predictive, preventive, personalized, and participatory. In this brave new world of heath care, we will be able to identify the earliest markers of disease and reverse them before the disease can develop. We will help patients change their behaviors and optimize their wellness. Together, these capabilities will prevent the development of many of the chronic conditions that ravage our collective health and cost our system billions every year. We will be able to quantify, characterize, and address a given patient’s risk of certain neurodegenerative and cardiovascular diseases. We will identify specific genes and gene combinations that dictate how an individual responds to different types of diet, exercise, supplements, and medications.
If knowledge is power, we are about to unlock more power over our health and wellness than the world has ever seen.
From Biology to Systems Biology
I’ve been a part of major changes over the past 50 years or so that have transformed biology and medicine from a science of studying single molecules or cell types to a cross-disciplinary, integrative, and systems-driven “big science” discipline. In the 1970s, when I was at Caltech, we invented four instruments that allowed one to read and write DNA and proteins, and we introduced high-throughput measurements and big data for biology. I participated in the Human Genome Project, which in 1990 began building a “parts list” of all genes and allowed us to begin applying systems approaches to biology and disease. In 2000, I cofounded the Institute for Systems Biology (ISB) to further develop the approaches of systems medicine and P4 medicine, creating a unique cross-disciplinary environment. In 2003, the first complete sequencing of the human genome gave biology and medicine a new and transformative source of big data. And systems biology brought a holistic approach to deciphering the complexities of biology and disease by combining and analyzing disparate data sources to achieve unprecedented levels of insight.
The Beginning of P4 Health Care
In 2014, I created a program in scientific wellness and persuaded 108 of my friends to allow my team at ISB to gather their data for a year to create for each a dense dynamic data cloud. We did whole genome sequencing, analyzed both their blood (clinical chemistries, proteins, metabolites) and their gut microbiome in great detail every 3 months, and had them take self-report measures — weight, diet, exercise, and the like. After analyses and integration, these data led to actionable possibilities to optimize wellness and avoid disease. We gathered data on me, too, of course. The results were dramatic for most. For myself, I can say that at 78, I feel like I’m 40, and when I’m 95, I expect to be feeling as I did at 50, all from changes that I’ve made as a result of understanding my data.
Based on data from this group and others that we have added since, we have made too many significant findings to list, but here are a few with implications that are very broad:
- We learned from blood tests that 90% of our initial cohort was low in vitamin D, so we recommended that all of these individuals take a standard supplement of 1000 IU/day. But it only worked on a few of them. From the genomic data, we learned that there are three different genes, each with two variants, which in the variant form block uptake of vitamin D. People with one or more of those variants needed more than the standard recommended dose, and if they had all three, they might need 10,000 to 20,000 IU to even begin to move into the normal range.
- It’s been known for a while that bipolar disease is associated with problems with specific neurotransmitters or calcium ion channels. We’ve studied 41 families with bipolar disease and have identified 125 gene variants that affect those functions. With this type of analysis, we can take a whole new approach to studying the complex genetic origins of many neuropsychological diseases.
- We believe that the standard model for clinical trials — taking 20,000 patients and assigning them to “matched” test and control groups — is fatally flawed because every individual is unique genetically and environmentally. Therefore, the groups aren’t matched at all. Indeed, in averaging these individuals together, we enormously increase the noise over the signal. In that process, we may miss discovering drugs that could dramatically help people with a specific genetic profile. In a P4-style clinical trial, we would determine the dense dynamic personal data clouds for 20,000 patients and could stratify them according to relevant clinical characteristics (e.g., response to an existing drug). We would be able to identify which patients respond to drugs by their molecular features and avoid testing drugs on known non-responders.
We have just scratched the surface of P4 medicine. As we accumulate and integrate new data on individuals, we identify new actionable possibilities, and there is no end in sight. We aspire to bring all individuals into their 90s with full mental and physical function.
The more data we have, the more discoveries we will make. We will amass some data through a commercial company, Arivale, founded in 2015 and backed by several venture capital firms, that already has more than 1,200 clients. We expect to have 10,000 within another year and more than a million by 2020. As we follow this large number of well people, we will begin to see wellness-to-disease transitions for all of the common diseases. We will use the tools of systems medicine to identify the biomarkers for these transitions and reverse disease at the earliest transition point.
In 2016, ISB affiliated with Providence St. Joseph Health, serving 3.3 million patients yearly, to become its research arm. Working with Providence’s patient population, we plan to apply this systems approach to specific areas, including scientific wellness, Alzheimer’s disease, breast cancer, and glioblastoma, an inevitably fatal brain cancer.
What Happens Now?
So, assuming our P4 concept is sound and will promote better health at lower cost in the way we expect, what has to occur to bring P4 medicine into full use in our system? I see the following issues that must be resolved:
Reducing the cost. Buying this type of dense dynamic data cloud development and analysis from Arivale, along with coaching to help apply the findings, is currently priced at about $4,000 — low enough to be within reach of people who are affluent rather than extremely wealthy but still not low enough. I’m confident that the cost of these assays will continue to drop as they become more widely used and the technology develops, but they must fall by two orders of magnitude to become accessible to all.
Convincing the patient. P4 medicine is such a radical change from traditional care that patients won’t necessarily understand it right away. We’re creating education modules about systems biology and P4 medicine so that students in high school biology classes can learn about it, and we can do similar things in lower grades as well. And for the general public, I’d love to see someone do a medical-mystery TV series that uses these techniques to solve interesting problems. With great writers and great actors, you could teach people an enormous amount.
Educating the physician. If incoming medical students received their own dense dynamic data clouds and learned how to analyze their own data, we could build an awareness of P4 medicine techniques and approaches into the entire medical curriculum. For physicians already in practice, we need to learn how to present them with the actionable data we discover, so that they in turn can present it to the patient and can gain an understanding of the value of this approach. Big data can be terrifying, and physicians need to feel assured that they don’t have to be computational biologists themselves in order to practice P4 medicine.
Addressing discrimination. When people discuss widespread genome sequencing, they frequently express a justifiable worry that the information will be used to deny care or insurance coverage, and that concern becomes even bigger with the dense dynamic data clouds that we are talking about. It is absolutely essential that we strengthen our laws to prohibit any type of discrimination based on health data because it will pose a major barrier to developing these powerful tools for wellness.
Ending the silos. So many research institutions have a territorial culture, and P4 medicine by its very nature is highly cross-disciplinary. In our work at Providence we have been delighted by the lack of silos, territories, and ownership. We had gone to several major medical schools and proposed to do just a small fraction of what we are doing at Providence, and overall we received a very cold reception (though one of our contacts said she wished it were possible to do this work at her institution). It may take a while to get there, but I think the only way to remove those silos is to demonstrate blazing success.