With advances in the scale and throughput of proteomics technology, the integration of predictive, preventive, and personalized medicine approaches into healthcare services may come sooner than anticipated. This follows a surge of research using ‘omics data to monitor the health of individuals, improve understanding of pathological mechanisms, and to discover novel biomarkers capable of identifying disease at its early stages.
One researcher who is working towards the development of these approaches using multi-omic data is Dr. Andrew Magis, Director of Data Science at the Institute for Systems Biology (ISB). In this interview, we learn more about how his team is using longitudinal proteomics measurements to identify early signals and transition mechanisms of diseases such as cancer. Magis also reveals how ISB is collaborating with Olink Proteomics to help improve the accessibility of its expanding proteomics platform and shares the intended impact of this technology on the future of healthcare.
Prior to his current role at ISB, Magis was the Director of Research at Arivale, a consumer wellness company that spun out of ISB in 2015, and closed in 2019. During this time, Arivale collected multi-omic data from its members with the goal of providing actionable recommendations to improve their health and reduce risk of disease. Blood, saliva, and stool samples were taken approximately every six months, and samples were biobanked and analyzed from more than 5,000 participants, 94 percent of whom granted consent to have their data used for research. “Arivale provided us with a unique dataset, as access to this many multi-omic datasets and samples from a population of ‘healthy’ individuals is not very common – samples are often only collected after people are diagnosed with a disease,” explains Magis. “This meant that if an individual developed a disease while they were enrolled in the program, we could analyze the samples that we'd already collected to see if we could find early signals of the disease.”
In a recent study, Magis’ team applied this approach to identify early signs of metastatic cancer. “We choose ten patients who had at least three blood draws that preceded their diagnosis, and analyzed around 1200 proteins in each of those samples,” he explains. By comparing this data against other participants in the program, the group identified specific proteins that persistently presented as outliers, constituting early signals of different cancers. “One protein in particular, the tumor marker CEACAM5, emerged as a significant outlier across three different metastatic cancer types, breast, lung, and pancreatic, in some cases more than two years prior to the diagnosis,” says Magis.
While CEACAM5 can be elevated for reasons other than cancer, Magis suggests that its coalescence with other markers could provide a valuable screening tool. “This is just one marker that could be combined with other protein markers or ‘omics data to provide confidence that a signal is real,” he explains. “Overall, we were able to see multiple protein markers changing months or years prior to the cancer diagnosis, and we believe that the convergence of this evidence made a strong case that prospectively could have justified increased surveillance.”
Magis’ group analyzes these proteins in patient blood plasma samples using Olink’s Proximity Extension Assay (PEA) high-multiplex immunoassay technology, which was chosen due to its specificity for target proteins, based on a dual-recognition methodology with antibody pairs linked to complementary oligos for DNA-based readout.
Since the time of the original study, Olink has announced a significant expansion of its protein biomarker discovery offering, supported by its high-throughput Explore 3072 platform with NGS readout. With this development, Olink will double the number of available protein biomarker targets, providing users with access to a library of over 3000 validated assays that provide broad coverage of all major biological pathways. “We’re really excited about the deployment of Olink's Explore platform,” says Magis. “Continuing the expansion in the sampling of more biological domains and relevant pathways will be incredibly valuable, both from a research perspective and if these technologies are ultimately incorporated at the clinic.”
He continues: “We see a bright future for this technology, so much so that ISB is becoming an Olink Explore 3072 provider. We have installed the Olink Explore platform in our Molecular & Cellular Core Facility for internal use and research collaborators, and will soon be offering the service for external users as well.”
Building upon this collaboration, ISB is also working with Olink to help contextualize data generated by the Explore platform using samples collected at Arivale. “Hundreds of these plasma samples will be run to build a database of ‘normal’ individuals across all of the different protein assays that have deployed in Explore 3072,” says Magis. “This database will allow researchers to compare the data in their studies to this population of healthy individuals to help identify outliers.” The data is to be made available to Olink customers in the future via Olink's Data Portal.
Magis is optimistic that with continued improvements in our ability to study major biological pathways, personalized and precision medicine approaches are set to become a central paradigm of future healthcare. “The technology is moving in a direction that we're all very excited about,” he says. “The future vision of all of this is that you get a small drop of blood or a regular draw at the doctor's office which is then able to provide a window into the state of your body.”
He continues: “Olink’s proteomics platform, and multi-omics more generally, is going to be extremely valuable, both in terms of early disease detection and monitoring the health of individuals and being able to understand the mechanisms of disease development in an unprecedented way.”
Magis also highlights the value of longitudinal multi-omics measurements to studying heterogeneity in health and disease phenotypes in future studies. “We remain very excited to collect longitudinal data in both healthy and disease cohorts,” he says. “We believe that studying healthy people is at least as important as studying sick people to understand how disease transitions occur and what it means to be well.”
1. Magis, A.T., Rappaport, N., Conomos, M.P. et al. Untargeted longitudinal analysis of a wellness cohort identifies markers of metastatic cancer years prior to diagnosis. Sci Rep 10, 16275 (2020). https://doi.org/10.1038/s41598-020-73451-z