The Human Plasma Proteome Draft of 2017: Building on the Human Plasma PeptideAtlas from Mass Spectrometry and Complementary Assays

Abstract

Human blood plasma provides a highly accessible window to the proteome of any individual in health and disease. Since its inception in 2002, the Human Proteome Organization's Human Plasma Proteome Project (HPPP) has been promoting advances in the study and understanding of the full protein complement of human plasma and on determining the abundance and modifications of its components. In 2017, we review the history of the HPPP and the advances of human plasma proteomics in general, including several recent achievements. We then present the latest 2017-04 build of Human Plasma PeptideAtlas, which yields ∼43 million peptide-spectrum matches and 122,730 distinct peptide sequences from 178 individual experiments at a 1% protein-level FDR globally across all experiments. Applying the latest Human Proteome Project Data Interpretation Guidelines, we catalog 3509 proteins that have at least two non-nested uniquely mapping peptides of nine amino acids or more and >1300 additional proteins with ambiguous evidence. We apply the same two-peptide guideline to historical PeptideAtlas builds going back to 2006 and examine the progress made in the past ten years in plasma proteome coverage. We also compare the distribution of proteins in historical PeptideAtlas builds in various RNA abundance and cellular localization categories. We then discuss advances in plasma proteomics based on targeted mass spectrometry as well as affinity assays, which during early 2017 target ∼2000 proteins. Finally, we describe considerations about sample handling and study design, concluding with an outlook for future advances in deciphering the human plasma proteome.

Keywords: Human Proteome Project; mass spectrometry; plasma; proteomics.

Figure 1

Number of distinct peptide sequences in the Human Plasma PeptideAtlas 2017-04 as a function of the total number of PSMs included (43 million). Each box represents one experiment. The width of the box corresponds to the total number of PSMs passing threshold. The total height of the boxes (blue+red) represents the cumulative number of distinct peptide sequences in the build (culminates at 122,730). The height of the blue boxes corresponds to the total number of distinct peptide sequences in each experiment.

Figure 2

Progression of the number of canonical proteins (using HPP Guidelines metrics) of historical Plasma PeptideAtlas builds. Builds since 2009 (in blue) do not contain N-glycosylation enriched datasets and were thresholded at a ~1% protein-level FDR prior to application of guideline 15 requirements. Builds prior to 2009 (in orange) contain N-glycosylation enriched datasets and were thresholded at a distinct peptide probability ≥ 0.9 prior to application of guideline 15 requirements.

Figure 3

A non-proportional 5-way Venn diagram depicting the overlap of canonical proteins in Human Plasma PeptideAtlas builds 2009 – 2017; 2013 and 2014 builds are combined because they are very similar. A core set of 984 proteins are found in all of these builds, with 1153 proteins added in 2013, 785 progressively added in 2015, and 573 progressively added in 2017. The number 714 in the lower corner represents the numbers of proteins unique to the 2006 and 2007 builds, which are now known to have quite high FDRs.

Figure 4

Relative distributions of estimated human plasma protein concentrations from four Human Plasma PeptideAtlas builds from 2010 to 2017. The mean of the distributions has moved markedly to include more proteins at concentrations below 1 ng/ml from 2010 to 2017.