NEW YORK (GenomeWeb) – The Human Proteome Organization gathered for its annual meeting this week in Taipei, Taiwan with researchers continuing their pursuit of the full human proteome.
According to an update published in the Journal of Proteome Research, HUPO's Human Proteome Project now counts 16,518 confidently identified proteins, 2,949 proteins that remain missing, and 588 proteins for which evidence of their existence is uncertain.
HUPO's Chromosome-Centric Human Proteome Project (C-HPP) has spearheaded the group's effort to characterize the full human proteome, and coverage has steadily increased since that initiative was launched in 2012. Progress has slowed in recently years, however, suggesting that the remaining proteins may prove more challenging to track down. Based on the contents of the NextProt database, researchers had identified 13,664 proteins as of 2012, 15,646 as of 2013, and 16,491 as of 2014.
One factor that affected the coverage numbers was a change in the project's inclusion criteria to increase the stringency of its guidelines in an effort to reduce false positives. Were the original guidelines still in place, an additional 485 proteins would be included as identified, which would bring the total to 17,003.
Of the 16,518 identified proteins, 14,658 were detected using mass spec, while 1,860 were detecting using non-mass spec methods like immunoassays.
The C-HPP is perhaps the centerpiece of HUPO's initiatives, but the organization's work spans a vast array of projects.
And, Swiss Federal Institute of Technology researcher Ruedi Aebersold noted during a presentation this week that numbers aren't everything. He suggested that as important as the overall number of proteins analyzed is the ability to put them in useful context — understanding, for instance, their relationships to one another and their functions and biology.
Aebersold helped launch HUPO's Biology/Disease-driven Human Proteome Project, proposing the initiative in 2011. Comprising more than a dozen smaller initiatives, the project is a complement to the C-HPP, coming at the challenge of mapping the human proteome from the perspective of the proteomic profiles of specific organs, diseases, and biological systems.
The meeting saw the launch of a new project, the Human Personal Omics Profiling (hPOP) study led by Stanford University researcher Michael Snyder.
Initially piloted at this year's US HUPO meeting in Boston, the hPOP study made its international HUPO debut this week, where Snyder recruited meeting participants as subjects for the project and as collaborators on data collection and analysis.
Snyder aims to collect samples from HUPO attendees at each year's conference and characterize their genomes, proteomes, and metabolomes, following hundreds or potentially thousands of subjects longitudinally over the course of years. Snyder told GenomeWeb this week that he collected samples from 109 people at this year's meeting.
HUPO is not typically a meeting with significant vendor news, but the week did see one new product release of note, Waters' SONAR data independent acquisition method, which the company said allows for data-independent acquisition (DIA) on UPLC time scales while improving selectivity by using a quadrupole to separate co-eluting precursor masses within the mass range being analyzed.
While conventional Swath-style experiments use instrument quadrupoles to select large m/z windows (in the range of 25 m/z) for fragmentation, in SONAR the quadrupole actively scans the selected mass window, further separating the precursors in that window.
"Due to the scanning nature of the quadrupole, we can deconvolute the precursor and product ions to approximately 1/20th of the quad window," David Heywood, Waters senior manager, omics business development, told GenomeWeb. He added that the method is intended for use on the company's Xevo G2-XS QTOF and will be available by the end of the year.
University of Washington researcher Michael MacCoss, who has done extensive work developing methods and software for DIA and other targeted mass spec techniques, told GenomeWeb that the technology behind Waters' SONAR looked "interesting," adding that it had similarities to work his group has been pursuing on using overlapping mass windows in DIA analysis.
MacCoss presented at HUPO this week on what the future of DIA might hold, highlighting as one example his team's MSX method, which uses the simultaneous measurement of multiple narrow mass windows to reduce precursor interference while retaining the method's breadth and speed.
Typically, DIA methods like Swath, select broad mass windows (in the range of 25 m/z) and fragment all the precursors in that window, which allows the instrument to collect MS/MS spectra on all the ions in that window.
These wide m/z fragmentation windows can create significant precursor interference, though, which leads to a loss of selectivity and reduces the number of proteins that can be identified and quantified.
Selectivity could be improved by narrowing fragmentation windows used, but this would limit the mass range measured in an experiment. For instance, at a mass spec acquisition speed of 10 Hz, a 20 m/z-wide window is required to sample a 400 m/z range every two seconds. Narrowing that fragmentation window would result in a proportional narrowing of the overall m/z range captured by the analysis.
In 2013, MacCoss and his colleagues put forth their MSX method, which addresses this problem by multiplexing smaller m/z windows drawn from different parts of the overall m/z range being analyzed. By simultaneously analyzing five different 4 m/z window, they can cover the equivalent of a 20 m/z-wide window but with the improved selectivity inherent in the smaller 4 m/z windows.
The original implementation of the technique suffered from several limitations, however. As MacCoss told GenomeWeb this week, the increased selectivity "came at the expense of sensitivity." Advances in mass spec instrumentation has largely solved this problem, he said, noting that the improved ion optics of Thermo Fisher Scientific's Orbitrap Fusion Lumos reduce the technique's sensitivity losses. The researchers originally developed the method on Thermo Fisher's Q Exactive instrument.
"Another limitation in the MSX implementation was that it was computationally expensive to demultiplex the data when there were lots of target peptides of interest," MacCoss said, noting that his team has since addressed this challenge, as well.
At this week's meeting he predicted for DIA's future the ability to analyze the entire m/z range with selectivity of better than 4 m/z, providing comprehensive proteomic data on samples while offering the dynamic range and sensitivity of narrowly targeted methods like parallel reaction monitoring that are typically used for measurements of a small number of proteins.