ORLANDO (GenomeWeb) – The annual meeting of the Human Proteome Organization held this week in Orlando saw the group continue its shift toward a more translational and clinical orientation as it closes in on the almost decade-long goal of identifying and characterizing at least one protein product for each human protein-coding gene.
That effort, the Human Proteome Project, had as of this week's meeting identified proteins from 17,470 of the roughly 20,000 human protein-coding genes, according to a presentation from the University of Michigan's Gilbert Omenn, one of the leaders of the project.
Omenn noted in an interview this week that while the project had appeared to plateau several years ago at around 16,000 proteins, the researchers had since made additional progress that encouraged them to carry on with the effort for several more years.
"We thought we were saturated a few years ago, and we almost declared victory," he said. "But we've managed to add more since then, so it appears worth going a little further."
That said, the project will almost certainly fall short of identifying proteins to every protein-coding gene given that "it is certain that some percentage will be undetectable," Omenn said. He cited the example of human olfactory receptors, a class of roughly 400 proteins, none of which HUPO researchers have yet detected using mass spec.
While the effort to catalog the human proteome may indeed plateau in the next few years, HUPO, like proteomics more broadly, is moving increasingly towards not just identifying discrete proteins but analyzing them in the context of specific isoforms, complexes, cells, and disease states.
One indication of this direction was the formal addition this year of a pathology component to the "pillars" of the HPP effort, which previously consisted of mass spectrometry, antibodies, and the bioinformatics knowledgebase.
"Pathology has been a part of the operation for a long time, but there was [more recently] a real groundswell of interest here from people who wanted more participation from real pathologists, not just researchers," Omenn said, noting that the organization this year provided several travel grants explicitly for pathologists to encourage their attendance.
"It's because of the increased interest [on the part of proteomics researchers] in clinical translation," he said. "What we hope it will do is integrate [pathologists] more than they have been up until now, to help guide people who don't know much about the regulatory world and CLIA."
"They want people to understand that you can't just develop a test and expect people to buy it, or, if it is paid for [by insurance], to use it," he added, "and that there are parameters for clinical chemistry and clinical pathology that are very demanding, very normal but outside the knowledge of many researchers. So it is a very important role and absolutely essential for translating to clinical practice."
The organization also this year launched a new initiative aimed at generating functional information for the roughly 10 percent of human proteins without any functional annotation, an effort that Omenn said will be key to including these proteins as part of larger systems biology efforts.
Northwestern University's Neil Kelleher suggested that the ongoing move within HUPO and proteomics generally towards exploring in more detail the nature of protein expression and function in specific biological systems presents an opportunity for his vision of a Cell-Based Human Proteome Project, which would aim to catalog the presence of intact proteoforms in specific cell types. Kelleher first presented the idea in a 2012 paper in the Journal of the American Society for Mass Spectrometry in which he proposed characterizing 4,000 distinct cell types at a depth of 250,000 intact proteoforms each, for a total of 1 billion proteoforms.
The idea has thus far gained relatively little traction outside the top-down proteomics community, likely due to the technical challenges that have limited uptake of top-down proteomics in general. Compared to bottom-up proteomics, top-down provides more comprehensive information on the different forms a given gene product can take depending on the particular post-translational modifications or genetic mutations affecting it. However, top-down is more technically difficult and less sensitive than conventional bottom-up approaches, and is, at this point, used by a relatively small group of expert labs.
However, Kelleher said he believed that with the HPP closing in on its goal, HUPO would be looking for its next major initiative.
"And I believe there is growing support for cell-based proteomics and cell-based top down as a way to improve our fundamental knowledge of proteins and protein complexes," he said.
He noted as an example of this interest a plenary presentation at this week's meeting by Stanford University Professor Stephen Quake detailing the work of the Human Cell Atlas, which, Kelleher said, provides both an example of how such an initiative might be undertaken. and a map of the various human cell types that could be used as a template for a cell-based proteome project.
That effort, which aims to generate single-cell RNA expression profiles from up to 100 million cells spanning a wide range of tissues and organs, has scored substantial funding from sources including the Chan Zuckerberg Initiative and the National Institutes of Health.
Kelleher promoted the idea in a presentation at this week's HUPO meeting. He suggested the organization is a potential vehicle for such an initiative but said that it might also be managed through an organization like the Consortium for Top Down Proteomics.
In any case, Kelleher acknowledged that additional technology development will be needed to make the project feasible but suggested that a large-scale initiative could spark that development.
He cited the role of the Human Genome Project in creating the incentives for the development of high-throughput sequencing technologies. He added that he expected the continued growth of biopharma as a major marker for mass spec vendors would continue to drive development of technologies for intact protein analysis.
"Biopharma has been one of the strongest drivers behind why great engineers and [mass spec vendor] managers have said, 'Yes, we need to develop better top-down technologies,'" he said.
Another emerging area that received heightened prominence at this week's meeting was single-cell proteomics, which received its own session, an indication of the field's progress in recent years.
Presentations during the session included work from Emma Lundberg at the KTH Royal Institute of Technology showing data from the Human Protein Atlas on the distribution of the human proteome at the subcellular level; from Akos Vegvari at the Karolinska Institute on a pair of tandem-mass tag-based approaches for single-cell analysis that were able to detect several hundred proteins inn HTC-116 cells and human monocytes; as well as single-cell or low-sample approaches from researchers at Northeastern University, Pacific Northwest National Laboratory, and the Max Planck Institute.
While HUPO is not typically a big meeting for vendor releases, there were several product introductions of note. Bruker released several method advances for its timsTOF Pro system including the incorporation of large-scale collision cross section values from peptides, which the company said provides an additional parameter for confirming peptide IDs, improving reproducibility across runs. The feature is being incorporated in the "match-between-runs" capability in the MaxQuant mass spec software developed by Juergen Cox's group at the Max Planck Institute of Biochemistry.
Biognosys and Evosep jointly announced a high-throughput workflow for clinical plasma proteomics that combines Biognosys' PQ500 plasma protein panel, which quantifies 500 proteins using DIA mass spec, with the Evosep One liquid chromatography system.