NEW YORK – A team led by researchers at Weill Cornell Medicine-Qatar has used Seer's Proteograph proteomics platform for mapping protein quantitative trait loci (pQTLs) in a cohort of 325 individuals.
The effort, detailed in a BioRxiv preprint last month, demonstrates the ability of Seer's technology to conduct relatively deep plasma proteomic analyses at fairly high throughput and indicates the potential of mass spectrometry for proteogenomic applications that have been dominated by affinity-based platforms so far.
Protein quantitative trait loci are links between genetic variants and plasma protein levels. They are typically characterized as either cis — meaning that the pQTL is located near the gene that encodes that protein — or trans, meaning it is located further away or on a different chromosome. A cis pQTL in many cases reflects the influence on a protein of the gene that codes for it, while trans pQTLs may reflect other phenomena, such as changes to other proteins that interact with the target protein or are in a signaling pathway with it. The hope is that pQTLs can help researchers map the connections between genetic variation and protein expression changes and, ultimately, disease.
Such experiments have become more feasible in recent years as the development of affinity-based platforms — primarily from Olink and SomaLogic — have allowed researchers to measure thousands of proteins in parallel in plasma samples from tens of thousands of individuals, bringing the depth and throughput of proteomic studies to a level where they can be meaningfully integrated with genomic datasets.
These affinity-based approaches have limitations. Many proteins exist in multiple forms, exhibiting alterations like amino acid variants or truncations or post-translational modifications, which can impact their function. It is often unclear or unknown with affinity platforms which particular form of a protein they are measuring. The existence of different proteoforms can, in fact, lead to situations where a pQTL mapped using two different platforms may differ depending on the platform used. In other cases, an affinity platform may not be able to measure a particular protein form due to, for instance, a modification or alteration that changes the epitope targeted by the affinity agent.
Mass spec-based proteomics, on the other hand, is better equipped to account for different proteoforms as it can detect features like variant peptides and post-translational modifications. Compared to affinity-based platforms, however, mass spec struggles with throughput and depth of coverage. While Olink's Explore platform measures around 3,000 proteins in plasma and SomaLogic's SomaScan platform measures about 7,000 proteins, high-throughput mass spec workflows typically top out at around 500 proteins.
Seer's Proteograph technology is intended to boost the depth of coverage of proteomics experiments while still allowing for relatively high throughput. The system uses nanoparticle-based enrichment of proteins and is based on the observation that when incubated in a biological sample, nanoparticles collect proteins, which form a "corona." Given this, nanoparticles can serve as an enrichment tool, allowing researchers to pull proteins out of a sample, which they can then identify and quantify using technologies like mass spec.
In their recent preprint, the researchers used the Proteograph system to identify more than 18,000 peptides from roughly 3,000 proteins across the 325 plasma samples they analyzed, providing a depth of coverage comparable to that of previous affinity-based efforts. From that data, they identified 184 protein-altering genetic variants that corresponded significantly with mass spec measurements of the matching variant peptide, associations that they termed mass spec protein altering variants, or MS-PAVs.
Comparing these MS-PAVs to pQTLs identified in prior affinity-based experiments, they found that 118 overlapped with a pQTL identified using SomaScan and 113 with a pQTL identified by Olink's platform.
Asim Siddiqui, Seer's senior VP of research and a coauthor on the study, said that while this was the first publication detailing the use of the company's platform for pQTL research, it demonstrated its potential for such work.
He suggested that mass spec might complement affinity-based experiments by validating or ruling out pQTLs identified in these studies. He also noted that the mass spec-based approach identified certain associations not picked up by affinity-based platforms. Specifically, 36 of the MS-PAVs detected had not been identified in previous affinity-based pQTL studies.
A mass spec-based approach "absolutely is useful," said Christopher Whelan, director of neuroscience data science at the Janssen Pharmaceutical Companies of Johnson & Johnson and chair of the UK Biobank Pharma Proteomics Project (PPP). Whelan was not involved in the Seer study but has led large affinity-based pQTL studies using the UK Biobank.
"Affinity-based platforms are allowing us to cast a very wide net. The mass spec-based platforms can help us parse out the proteins that this net has caught," he said.
"Overall, it is a very strong preprint," Whelan said. "[It] shows that there are certain proteins or pQTLs that the wider [affinity-based] nets sometimes miss, so you occasionally need to cast that smaller, finer net with mass spec to discover more complex associations."
"Mass spec is still seen as the gold standard for proteomics, so in one sense, it serves as a validation tool for the affinity-based proteogenomic findings," he said. "In another sense, it can help us find associations that aren't possible to find with the affinity-based techniques. Post-translational modifications and proteoforms are probably the biggest use cases."
He cited the example of Alzheimer's disease, where modified proteins like phosphorylated versions of tau are among the most prominent biomarkers for the condition.
Mass spec still struggles to match the throughput of affinity-based platforms, however, particularly when analyzing plasma, which limits its ability to look at the effects of less common variants.
One thing to note about the study is that it applied a minor allele frequency cutoff of 10 percent, Whelan said. "But that is going to be unavoidable with a sample size of 320. Anything lower than 10 percent would produce lots of false positives."
He suggested that this throughput limitation "highlights how mass spec-based proteogenomics is still a somewhat specialized field."
To identify pQTLs involving rare variants "you really need sample sizes of thousands if not tens of thousands to detect a reliable association," he said. "I think [mass spec] currently works best as a follow-up tool or validation tool for the larger affinity-based proteogenomic studies."
"Certainly our goal at Seer is to enable studies of that [larger] size," Siddiqui said. He noted that diagnostics firm PrognomiQ — a spinout of Seer — is planning a roughly 15,000-sample study using the Seer platform.
"So certainly studies of tens of thousands of samples are not beyond the capabilities of our technology platform," he said.
Whelan said that the UK Biobank PPP is currently looking to implement mass spec proteomics as part of its work, which to date has relied on affinity-based platforms.
"We might soon be able to employ mass spec in 55,000 people like we did with an affinity-based approach, but we are first exploring ways to implement mass spec in over 1,000 samples as a proof of concept," he said.