NEW YORK – Seer remains on track for a broad commercial release for its Proteograph proteomics platform next year with several of its early-access users recently presenting data generated using the system.
At the American Society for Mass Spectrometry annual meeting earlier this month in Philadelphia, researchers from the Oregon Health & Science University, the Broad Institute, Sanford Burnham Prebys Medical Discovery Institute, and proteomics software company Protein Metrics made poster presentations detailing their use of Seer's Proteograph technology.
Seer's Proteograph system uses nanoparticle-based enrichment of proteins in samples like human plasma to enable deeper proteomic discovery experiments at high throughput. The technology 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.
From a proteomic profiling perspective, it is notable that nanoparticles appear to collect proteins from across a large portion of the dynamic range present in samples like plasma. This means they could allow researchers to analyze low-abundance proteins without using time-consuming and expensive steps like protein depletion and fractionation that are required for mass spec-based experiments that aim to go deep into the proteome.
In a paper published last year in Nature Communications, Seer employees and affiliated researchers used the platform to analyze 141 plasma samples from non-small cell lung cancer patients and healthy controls, identifying an average of 1,664 proteins per subject and more than 2,000 proteins across all subjects combined.
Seer researchers also presented a poster at this year's American Society of Human Genetics annual meeting in October, in which they used the platform for a proteogenomic analysis of a subset of the lung cancer samples from the 2020 Nature Communications study. They sequenced the exomes of 29 subjects from that study and used that data to create mass spec search libraries for each individual. Using those databases, they identified 464 peptide variants in the 29 samples, with those variants mapping to seven of the 16 proteins the researchers had previously identified as differing in expression between lung cancer cases and controls.
Margaret Donovan, product marketing manager of bioinformatics at Seer and first author on the study, said that she and her colleagues hypothesized "that potentially it is alternative splicing that is contributing to these differences in amino acid sequences," though she noted this would need to be validated in additional experiments.
The research presented at the recent ASMS meeting further indicates that Seer's technology can enable deeper coverage of the plasma proteome than has been possible to date with workflows offering a similar level of throughput.
The Broad researchers used the system to profile the plasma proteome of three patients undergoing planned myocardial infarction, looking at blood samples taken at five different time points, detecting an average of 2,200 proteins per sample and more than 4,000 proteins across all 15 samples. The work, which was done on a Thermo Fisher Scientific Exploris mass spectrometer using 1 hour LC gradients, reanalyzed three sets of patient samples analyzed in a 2017 experiment by several of the same Broad researchers that presented one of the deepest analyses of the plasma proteome to date, quantifying more than 3,500 proteins.
That effort, however, relied on extensive fractionation of the patient samples, requiring more than a week of sample preparation work and 120 mass spec runs. Additionally it used a depletion column, the GenWay SuperMix, that Steven Carr, director of proteomics at the Broad and an author on both studies, said was no longer available, which would hinder the ability of future studies to achieve the same depth of coverage via depletion.
By contrast, the work presented by the researchers this month at ASMS — though not reaching as deep into the proteome as the original study — required seven hours of automated sample prep on the Proteograph system followed by 80 LC-MS runs. The researchers detected clinically relevant proteins including cardiac troponins as well as more than 150 of the roughly 300 regulated proteins detected in the original experiment and new potential markers of planned myocardial infarction. Additionally, compared to the earlier study, the Seer work detected the cardiac troponins at earlier time points and detected more peptides per cardiac troponin protein.
The OHSU team showed data from a small pilot study analyzing 32 plasma samples collected from prostate cancer patients, half with low-grade disease and half with high-grade disease. Using the Proteograph system along with a Bruker timsTOF Pro mass spectrometer, the researchers identified an average of 947 proteins in the high-grade samples and 960 in the low-grade samples and 1,545 proteins total. The OHSU researchers are now following up with a highly scaled prostate cancer study of 500 to 1,000 specimens with a clearly defined clinical endpoint.
The Protein Metrics team focused on glycoprotein analysis, finding that Seer's nanoparticles provided effective enrichment of plasma glycoproteins and could allow for the study of these proteins without the need for additional glycopeptide-specific enrichment.
The Sanford Burnham Prebys researchers, meanwhile, combined the Proteograph system with tandem mass tag (TMT) multiplexing, using the 16plex TMTpro reagents and fractionation into 24 fractions following processing with Proteograph. They then ran the fractions using a 100-minute LC gradient on an Orbitrap Fusion Lumos outfitted with a FAIMS Pro system. Using this workflow, which could run eight samples per day, the researchers detected a total of 2,785 protein groups across nine orders of magnitude, including low abundance analytes like cytokines.
On Seer's recent Q3 earnings call, Omead Ostadan, the company's president and chief operating officer, highlighted the depth of coverage exhibited by the Sanford Burnham Prebys work, noting that the researchers were able to identify proteins, like cytokines, present in the picogram per milliliter range. That represents a significant improvement in sensitivity compared to typical mass spec-based plasma proteomic experiments.
Daniel Hornburg, senior director at Seer, said that in internal work the company is able to get four times the number of protein IDs using its system compared to standard plasma proteomic workflows when analyzing undepleted plasma. When comparing to analysis of depleted plasma, the researchers typically get around a threefold improvement in protein identifications, he said. He added that with its internal workflows employing the five nanoparticles that make up the existing Proteograph system, the company is currently able to measure more than 3,000 proteins per plasma sample.
Beyond Seer, a number of companies are working to improve plasma proteomic analyses, aiming for the combination of depth and throughput that has long eluded the field. Because Seer is fundamentally a sample prep company, its technology may be compatible with and additive to many of these workflows, as well as to other emerging proteomic technologies.
This month, for instance, Swiss proteomics firm Biognosys began offering a blood-biomarker discovery service that the company said will quantify up to 3,000 proteins per sample.
Proteome Sciences, meanwhile, has begun running large-scale experiments using its TMTcalibrator tool, which can quantify more than 5,000 proteins in plasma, though the throughput for this workflow is relatively low.
SomaLogic's aptamer-based SomaScan platform, meanwhile, can measure 7,000 proteins in plasma. Meanwhile, Olink's Explore platform currently measures 1,500 proteins in plasma, and the company plans by the end of the year to start shipping a new version of the platform that can measure 3,000 proteins per sample.