Plasma Proteomics With Professor Joshua Coon

Human plasma has long been considered a key target for proteomic analysis, especially as medicine moves toward more personalized approaches. Plasma – the liquid component of blood – contains thousands of proteins. While many of these proteins function in the circulatory system, others are expressed in response to stimuli, such as the manifestation of a disease. Plasma can therefore be likened to the body’s “sink”, representing a rich source of blood biomarkers that could be used to assess healthy or diseased states in a minimally invasive manner.

However, the infamous dynamic range challenge of the plasma proteome has significantly limited its study. While some proteins – such as albumin – are highly abundant in plasma, others are expressed at incredibly low concentrations. This enormous range poses a significant analytical challenge – one that has caused many researchers to step away from plasma proteomics.

“Researchers have recognized the potential applications and how useful plasma could be for a long time. But the challenge has always been the dynamic range in that sample,” Professor Joshua Coon told Technology Networks. “For the first 10 years of its existence, my lab avoided plasma proteomics because it was just too difficult to get good depth.”

At the University of Wisconsin-Madison and the Morgridge Institute for Research, the Coon Laboratories innovate new mass spectrometry (MS)-based methods and apply these techniques to advance biomedical research. He is also on the scientific advisory board of Seer Inc, a consultant for Thermo Fisher Scientific Inc and a founder of CeleramAb Inc.

“It’s only within the last six or seven years that we have tried to get research moving in plasma proteomics,” Coon said. “It requires two things to do it well: you need a really good liquid chromatography-mass spectrometry (LC-MS) system, and you need a way to prepare your sample to counter the dynamic range.”

Until recently, researchers simply lacked such tools and methods. However, the commercialization of increasingly sensitive MS technologies and enrichment methods has enabled a “revival” of plasma proteomics.

“Now, it's exciting because we can tackle discovery projects in plasma proteomics that just hadn’t been possible before,” Coon said.

He cites a recent example: “When the COVID-19 pandemic hit, like most labs, we were sent home. But I had a collaborator in New York, Professor Ariel Jaitovich – and since New York was one of the hardest-hit areas in the US, he was collecting blood samples from patients presenting with COVID-19.”

By May 2020, Professor Jaitovich had collected over 100 blood samples from patients with COVID-19 of varying severity, alongside other metadata. He contacted Coon to ask whether his laboratory would conduct a plasma proteomics analysis. “We did – and we also ran metabolite and lipid analyses,” Coon said. “At that time, we had decent mass spectrometers, but no enrichment method. Still, we were able to quantify around 300 to 400 proteins and identified some that were predictive of COVID-19 severity.” The work is published in Cell Systems.

Jaitovich has continued working on COVID-19 samples, focusing on long COVID. The cohort has now grown to ~500 samples. “We’re currently finishing a project using Seer’s Nanoparticle Technology combined with the Orbitrap Astral – two technologies we didn’t have just five years ago. Now we’re quantifying ~6,500 proteins per sample and identifying significantly improved signatures of both acute COVID-19 severity and long COVID,” Coon said.

“It’s been incredible to see how dramatically our ability to deeply profile the plasma proteome has advanced in just a few years."

Mass spectrometry versus affinity-based plasma proteomics

Coon’s laboratory specializes in using MS techniques to study the proteome.

MS proteomics is considered the gold standard approach for plasma biomarker discovery and protein profiling, particularly since the advent of data-independent analysis methods. In recent years, however, new and potentially complementary approaches have emerged, expanding the toolkit further.

Historically, affinity-based proteomics methods have been limited by the availability and specificity of antibodies, restricting the range of analytes that could be detected. However, newly commercialized versions of these technologies boast the ability to measure hundreds to thousands of proteins simultaneously.

Studies comparing MS and affinity-based methods are beginning to emerge, sparking some debate over whether one is superior – or whether both have distinct, complementary roles in proteomics.

Coon and colleagues recently published a paper in the Journal of Proteome Research that provides a rigorous comparison of six plasma proteomic technologies specifically. This included Olink Explore® HT and five MS-based methods; neat (control or nonenriched plasma preparation), acid (perchloric acid method), Mag-Net, PreOmics® ENRICHplus (pre-released kit) and Seer Proteograph XT.

Each method was assessed for achievable proteomic depth, reproducibility, linearity of measurement, robustness to lipid interference, global limit of detection and quantification and use in a differential expression study.

The Science and Future of Biomarker Discovery

This infographic explores key methods for identifying biomarkers in addition to highlighting the growing role of AI and machine learning in managing complex datasets and accelerating meaningful insights.
View Infographic

Advertisement

“I had already been using Seer’s technology and was curious to explore others, particularly affinity-based methods, to see how well they performed,” Coon said. “Our goal was simply to evaluate what each technology did best — to understand their respective figures of merit. We decided to include Olink in the study because of its strong reputation, proteomic depth and quick turnaround,” he continued.

“We sent the samples to an Olink core laboratory for analysis, while all MS-based experiments were conducted in-house. For each MS method, we coordinated with the relevant academic lab or company to ensure we followed best practices,” Coon said.

Collectively, the researchers conducted 618 LC-MS/MS experiments, and 93 Olink Explore HT analyses. Their key findings were:

• Sample preparation method critically impacts MS plasma performance, especially in terms of proteome depth.
• Among MS-based methods, Seer’s Proteograph technology outperformed all others. It offers nearly double the depth and superior technical metrics compared to Olink.
• MS-based approaches support unbiased discovery, allowing detection of post-translational modifications, protein variants and non-human proteins (e.g., bacterial), whereas the Olink strategy is “best described as a targeted proteomic technology”.
• Olink offers high throughput thanks to its next-generation sequencing components and is suitable for large-scale cohorts (up to 50,000 samples).
• There was minimal overlap between proteins detected by MS and Olink, highlighting their complementarity.

"The study was really eye-opening for me. I had a solid understanding of the strengths and weaknesses of MS methods, but very little experience with affinity-based approaches,” Coon said.

“On paper, they look incredibly promising – and after using them, I do believe they’re ideal for large-scale, targeted analysis. That said, if your goal is true discovery, I still think MS remains the leading approach," he continued.

At the 2024 Human Proteome Organization World Congress in Dresden, Germany, attended by Technology Networks, a consensus among researchers became apparent: no single tool will suffice for the future of plasma proteomics. Instead, the focus should be on expanding both the technological toolbox and the skillsets of researchers to enable the effective use of multiple complementary approaches.

"I think that’s true,” Coon said. “Honestly, it wasn’t my view before we conducted the study. When you read the literature on affinity-based methods, the message is often, ‘we can detect thousands of proteins at scale’, but in reality, many of those proteins aren’t actually present in plasma, so you only detect them in rare cases.”

“That said, these methods do scale incredibly well, which is something MS still struggles with. I do think MS technologies will improve in that regard, but it won’t happen overnight. On the other side, I hope we’ll see affinity-based platforms expand their ability to detect more of the proteins that are truly present in the plasma proteome.”

Panel Discussion: Environmental Pollutants

Join leading experts for a dynamic panel discussion exploring innovative analytical techniques used in environmental monitoring.
View Webinar

Advertisement

Coon believes his publication is a “solid starting point” for comprehensive comparison studies. The team is already planning to expand it to include SomaLogic’s SOMAscan® technology.

Plasma proteomics: Challenges and future perspectives

Despite significant technological advances, there remain several barriers when translating lab-based plasma proteomics research to the clinical space.

Coon emphasized how even with relatively small studies – i.e., under 600 samples – he has experienced issues with the number of pre-analytical factors that can impact sample quality. “These factors aren’t always obvious going in [to the study], but once you analyze the data, the effects can be dramatic – and they can significantly limit the conclusions you’re able to draw,” he said.

Plasma proteomics could have a significant impact on the future of personalized medicine. Credit: Stock.

“Take one example from the aforementioned COVID-19 study. The samples were collected over five years, and honestly, it was one of the best-designed studies I’ve worked on. The same physician took the blood samples each time, and either he or a colleague prepped and froze the plasma right away. That consistency minimized variation in sample handling, which can have a huge impact,” Coon continued. “But in our last dataset, when we looked at the data informatically, we noticed something odd with our figures. At first, the team couldn’t identify anything that had changed over the dates the samples were collected. Then, a couple of days later, they realized that it was the date when they had ordered a new supply of tubes – from the same vendor and supposedly the same product. Yet, this tube resupply appeared to cause a significant shift in the plasma proteome.”

In a clinical setting, such issues could have disastrous effects. Until recently, scientists didn’t have the analytical depth to even notice these issues. “But now we do,” Coon said, “And I think researchers need to be cautious, especially when going back to analyze biobank plasma that’s 20 years old for clinical research purposes.”

Pre-analytical variation is a “fascinating area”, Coon said, noting that he feels as a comprehensive study on what it can look like, and why it happens, is lacking. “It deserves more attention and support. Hopefully, someone will take it on, because it really is a concern for the field,” he said.

Looking to the future of plasma proteomics, Coon believes the next big step forward for MS proteomics is going to be throughput. “We need to process more samples faster to tackle large-scale studies, and I think we’ll see major innovation there soon,” he said.

“For affinity-based methods, I’d like to see broader protein coverage. Our comparison study showed that MS detects many proteins that affinity platforms currently miss. There’s a real opportunity to combine these approaches – merging data streams and using overlapping detections to improve accuracy,” Coon added. That complementary potential hasn’t been fully tapped yet, but it could be transformative, he concluded.