Skip to main content
Premium Trial:

Request an Annual Quote

Snyder Lab Combines Microsampling, Multiomics for Health Tracking, Biomarker Discovery

Premium

NEW YORK – Stanford University researchers have combined frequent microsampling with multiomic measurements and wearable sensors to produce personalized profiles tracking changes in thousands of biological molecules throughout the course of daily activity.

Detailed in a study published this week in Nature Biomedical Engineering, the work points toward more convenient processes for conventional clinical testing as well as new approaches to health monitoring and biomarker discovery, said Michael Snyder, professor of genetics at Stanford and senior author of the paper.

Snyder and his team have also recently launched two companies, Iollo and RHTM, to commercialize microsampling-based health monitoring.

Snyder has long been interested in combining multiomic measurements with longitudinal sampling and has launched a number of studies in this area. Perhaps most famously, his lab published a study in 2012 analyzing his genome, transcriptome, proteome, metabolome, and autoantibody profiles — dubbed the "Snyderome" — over a period of one to two years, during which time he was able to detect and subsequently manage his early development of type 2 diabetes.

In the most recent study, Snyder and his colleagues set out to evaluate the utility of various microsampling approaches as well as the stability of different analytes in microsamples under various conditions.

While dried blood spots have commonly been used for microsampling, it can be difficult to collect consistent sample volumes, and Snyder said that his lab ran into reproducibility issues with this approach. This led the researchers to explore newer microsampling devices that collect a defined amount of blood from fingerpricks. Ultimately, they settled on the Mitra collection device from microsampling firm Neoteryx, a subsidiary of Australian life science firm Trajan.

They then investigated the stability of a set of proteins, metabolites, and lipids in microsamples stored at a variety of temperatures and over different durations. Proteins proved the most stable, with 17 of 128 impacted by storage conditions or duration. Metabolites were less stable, with more than a quarter of the 1,462 analytes measured impacted, while more than half of the 776 lipids measured were affected by storage conditions.

Nonetheless, Snyder said, the results indicated that a substantial number of molecules are stable enough to make for useful measurements.

The researchers also looked at the correlation between measurements made using microsampling and those made using venous blood draws, finding that most classes of molecules were very similar between the two.

Having determined which molecules they could measure effectively, Snyder and his colleagues then used their microsampling workflow to track multiomic changes in 28 individuals after consumption of an Ensure shake. To do so, they had participants collect and mail in samples prior to drinking the shake and then at 30, 60, 120, and 240 minutes afterwards. Using a combination of mass spectrometry and Luminex-based immunoassays, they measured 560 metabolites, 155 lipids, and 54 cytokines or hormones from each individual across the five time points. Analyzing those measurements, they tracked the behavior of various biological pathways following consumption of the shake over time and between participants.

Different participants "responded very differently" to the shake, Snyder said. For instance, in some it reduced their levels of inflammatory markers and in others it raised them. He noted that researchers have long observed that different people can have different reactions to the same foods, suggesting that the frequent multiomic measurements could help tease out the biology underpinning these different responses.

Likewise, the approach may prove useful for studying individuals' responses to a variety of other activities or interventions.

"We don't really do much of that in medicine today, except in very specialized cases like stress echocardiograms," he said. "I think this could open up a lot of opportunities, to make abnormal measurements where you could learn something pretty important about people's health."

The researchers also made multiomic measurements on a single individual sampled intensively (98 total samples) over the course of a week. They combined those measurements with wearables, step count, food log, exercise, and continuous glucose monitoring data collected over the same time period. Using that information, they explored a number of questions including the circadian rhythms of the molecular measurements and their correlation with wearables data.

To make such testing commercially available, Snyder cofounded two recently launched companies — Iollo and RTHM — that sell microsampling-based health monitoring. Iollo, which is led by his former postdoc Daniel Gomari, sells a subscription service through which customers can get measurements of 500 metabolites with frequency ranging from once per year (for $276) to nine times per year (for $1,908). Additional services, which vary depending on the package, include integration of the metabolomic data with data from wearables and diet tracking apps and, at the higher end, with personalized sleep, diet, and exercise plans and 30-minute consultations with scientists from the company to go through the results.

The second company, RTHM, aims to use regular multiomic measurements to help manage a variety of chronic conditions. Ryan Kellogg, a former Snyder postdoc and co-author of the Nature Biomedical Engineering paper, is RTHM's CEO.

Leigh Anderson, whose protein analysis firm SISCAPA Assay Technologies offers a similar, though more narrowly focused microsampling product through its LongitudeDx business, said that Snyder and his colleagues had done "a tour de force" in terms of putting together a multiomic, multisampling workflow, but he questioned the extent to which they had shown such an approach could generate actionable medical information.

While the variation in individual analyte levels is real, "we have known that people are extremely variable in this sense for 50, 60 years," he said. "What you want to see is a demonstration that you can aggregate all this stuff, and it can give you a clear answer about something that you can really use."

Anderson said that although he spent years working in proteomics, which "is all about measuring lots of things," he has more recently come to suspect that making longitudinal measurements of a smaller number of relatively well-characterized markers might be a more informative approach.

"I feel like we have seen enough demonstrations of the idea that you are going to need big data and some magic interpretive to come up with the simple answer you can actually use," he said. "Machine learning is a powerful thing, and there are lots of interesting observations that come from looking at more than a few analytes."

"But," Anderson added, "I'm increasingly convinced that what we really want to do is understand a multiplicity, but a small multiplicity, of things and figure out how to choose the right ones and put that together longitudinally and look at those changes in people."

LongitudeDx's panel is primarily focused on inflammation markers, he said, adding that the company is in the middle of running between 10,000 and 15,000 samples that it hopes to publish on later this year.

While Anderson might question the value of extensive analyte panels, Snyder said his lab aims to add more molecules to its measurements to generate even deeper profiles. He noted that while some of the markers have known implications for people's health and lifestyle, discovery of new markers is a main goal of the work. The data generated for Iollo and RHTM customers is being used in that effort, he added.

"We hope to discover new markers for mental health, chronic diseases, and other things," Snyder said, citing diabetes as an example.

"Within type 2 diabetes are many different conditions, and we are trying to tease that apart right now," he said. "And I think that is going to be true for all complex conditions, and in a sense for all behavior. We have very different reactions to food, different reactions to all kinds of things. I think the more we can recognize that and define it biochemically, the more we can act on it appropriately."