MOUNTAIN VIEW, Calif. (GenomeWeb) – In a panel at this week's Personalized Medicine World Conference, executives from several companies gathered with researchers to discuss why they believe the moment has arrived for genomics to move into the promotion of general lifetime wellness, rather than being relegated solely to disease detection and treatment.
The discussion featured Stanford's Mike Snyder, who updated PMWC attendees on his comprehensive 'omics personal health research. Snyder — known for being both investigator and subject of a study of himself, the so-called "Snyderome," published in 2012 — has expanded the protocol he applied to his own biology to a cohort of about 100 subjects.
Also present were the leaders of two companies that have recently turned their focus toward the clinic, Lee Hood and Clayton Lewis of Arivale, and Brad Perkins of Human Longevity. The two firms have taken similar, but slightly different approaches to the goal of amassing comprehensive 'omics and, clinical data, and longitudinal data in order help their customers change their behaviors and nutrition regimens, manage their medications, and other interventions aimed at improving wellness and disease prevention.
According to Snyder it was this ability to combine genomics, blood chemistry monitoring, and microbiome analysis that allowed for the identification of his elevated risk of diabetes, as well as the early detection of his transition to an elevated blood sugar, and the linkage of that shift to a respiratory viral infection.
Making these same kind of connections as a way of validating hypotheses — for example, the connection between viral infection and diabetes inception — is a central goal in Snyder's expanded study, he said at the conference.
"The goal for project is to look at how people change over time — what they look like during health, and how things change as people acquire disease or go through other perturbations," he said. "We presume there are causal relationships, but finding out how that works and if you can manipulate it will be hugely important."
Snyder's analysis of himself has been going on for six years. In the expanded study, he and his colleagues have recruited about 100 people, who they have been monitoring for up to three years. The group has collected about 1000 time-point samples for these individuals, and is in the process of analyzing the data.
Sharing some tidbits, Snyder said that genome sequencing in his subjects has revealed clinically significant information in three individuals, including one who had rare mutation that appears to explain a family history of heart attacks. The study has confirmed that, even at baseline, every individual is clearly different, he added. For example, the study subjects have shown striking differences in their cytokines, with a presumed corresponding difference in their ability to fight infection. Microbiomes have been likewise unique, Snyder said, as well as robust from individual to individual.
Hood and Lewis gave an update on Arivale, which they launched commercially on the heels of a research study called the 100K Wellness Project, initially presented by Hood two years ago at the 2014 PMWC.
Hood shared some initial results from the first year pilot phase of that study in May 2015 at the University of Alberta, in which he and colleagues at the Institute for Systems Biology sequenced the whole genomes of about 107 "pioneer" individuals, who also submitted blood, urine, saliva, and fecal samples every three months for metabolite, proteomic, and gut microbiome testing, tracked their fitness using wearable devices, shared information on their lifestyle habits, and received coaching for better wellness based on their data.
Hood and his colleagues just submitted their first detailed report on results from the 107-subject pilot to an academic journal. Though he did not go into much significant detail about the findings, Hood said that "the approach has been very much like the Hubble telescope to the dark matter of the universe. We looked at data never before seen in human studies and made correlations never made before."
As part of the study, researchers were able to correlate genomic and other data with what Hood called health "transitions," for example, a well person becoming sick, or conversely, becoming even healthier.
About half of the overall cohort, 53 individuals, entered the study in a pre-diabetic state, but were all able to move their blood sugar status toward the normal range. While Hood did not discuss the relative contribution of genomics in the case of the pre-diabetic patients in the study, he said there have been cases where it was clear that there is indeed an added value to integrating genomic information in addition to traditional chemical blood test data, such as vitamin D levels.
In the 100K wellness pilot group, 92 individuals — including Hood himself — had very low vitamin D. The researchers were able to identify six variants in three genes that block the uptake of Vitamin D in members of the cohort, including in Hood.
"I started taking 1,000 units of vitamin D with no change in my levels, [then increased to 5,000] up to 10,000 units and now I'm slowly inching up to normal," he said. "It really illustrates that you can't just know the clinical chemistry, you need to know the genomics."
Commercialization
Unlike Snyder's academic study, Arivale and Human Longevity are now shifting from research to offering paying customers a service based on these approaches.
In June 2015, Arivale announced that it had raised $36 million in a Series B funding round, with plans to begin offering a commercial service that mirrors what was done in the trial — initially in Seattle and San Francisco, and then expanding into other markets in 2016. As Arivale moves forward into its commercial business, Hood has said that all participants from the 100K pilot study have signed on to become customers — the service costs $2,000 per year.
When it launched in July in the Seattle area, Lewis said Arivale saw almost 1000 individuals sign up. The company is also preparing to launch in Q2 with two or three major corporations on the west coast that want to make Arivale available to their employees, he said.
Human Longevity opened its first direct medical services platform last October, which, somewhat similarly to Arivale, aims to develop individualized risk assessments for early detection of early causes of mortality, albeit at a much higher $25,000 to $50,000 cost.
In addition to its initial La Jolla branch, the company also plans to open more of these "Human Longevity Health Nucleus centers" in South Africa and the UK.
In addition to whole genome sequencing and phenotypic assessments of various types, Perkins said that Human Longevity, unlike Arivale, is also testing its customers using advanced imaging technology. The company currently has one MRI machine in its La Jolla facility, and is adding another along with CT imaging capabilities. Findings from the company's analyses are returned to patients via a 3D avatar, which can also be used in video gaming.
Reflecting on the first 60 or 70 patients analyzed in this new clinical context, Perkins said there have been frequent findings that represent "new medical intelligence" for individuals who thought they were healthy going into the process, including the detection of polycystic kidney disease, ventricular hypertrophy, and a "really interesting" inborn error of metabolism in one subject. About 30 percent of the company's early clients have had some similar finding, he added.
Aside from addressing individual customers, both Arivale, and Human Longevity also have a parallel goal of collecting genome sequences and other data for very large cohorts of patients, something that likely will be necessary to further develop and validate the utility of their approaches in actually informing or affecting health and longevity.
Hood's 100K study aims to track 100,000 subjects longitudinally over ten years, but he did not provide an update at the meeting on how many subjects have been recruited so far.
In contrast, Perkins said that Human Longevity expects to reach a run rate of 100,000 whole human genome sequences by the end of this calendar year based on its current scaling plans. It has completed 20,000 so far. Over the last 20 months, he said the company has grown to 260 employees in three locations, including what the firm believes is the largest whole-genome sequencing center in the world. Perkins said that Human Longevity's organizing principle is to "predict everything that can be predicted from the whole human genome sequence."
"We think this is the principle engine to revolutionize the practice of medicine both in our current single disease model, but also providing a bridge to models of medicine that address the basic biology of aging," he added.
During his talk, Perkins shared an abstract from a study that the company recently submitted to a major academic journal on data from 10,000 whole genome sequences, what he said is an order of magnitude greater than any other publication in terms of high coverage human WGS data. The company found that rare variants were extraordinarily common in this 10,000-person cohort. The study identified 150 million rare variants, he said, more than half of them never seen before.
Moreover, each individual in the study contributed about 9,500 variants that were only seen a single time in the 10,000 genomes, according to Perkins. "We think that in 100,000 genomes we will identify more than 500 million rare variants, and we think — in contrast to GWAS and Mendelian disease — that this is where the action is going to be in terms of prediction of human phenotypes," he said.
Finally, Perkins added, the company has been studying the prediction of traits from human sequences, including experiments in a 1,000-person subcohort on predicting voice and sex, as well as facial structure and appearance. He revealed the progress the company has made in facial recognition with what is now the eighth generation of algorithms it has developed to create a photo-quality facial representation from genetic data alone.
"It shouldn’t be a surprise to anyone that we can actually do this with a fair amount of accuracy," he said, "which has important implications for the de-identification of whole human genome sequences."