Objective:

To map the diversity of protein forms (proteoforms) in human plasma and examine how proteoform diversity relates to measurements from affinity-based proteomic platforms.

Approach:

• Sample preparation: Researchers depleted 14 highly abundant plasma proteins and separated intact proteins by size and electrical charge using three methods.
• Proteomic analysis: They analyzed 3,114 fractions with data-independent mass spectrometry and computationally reconstructed peptide signals across fractions.
• Confidence assessment: Proteoforms detected in at least three of four replicate gel experiments were classified as high confidence.

Key Findings:

• The dataset contained 179,049 potential proteoforms across 5,077 protein groups; 44,083 proteoforms from 2,089 groups met the high-confidence criterion.
• The median was 11 high-confidence proteoforms per protein.
• Proteoform diversity was greatest among abundant proteins and proteins involved in nutrient transport, blood clotting, and the complement system.
• The most abundant form accounted for less than half of a protein’s total measured peptide signal.
• Olink and SomaScan measurements generally agreed more closely for relatively abundant proteins with low or moderate proteoform diversity. Poor agreement was most often associated with low protein concentrations.

Interpretation:

The authors note that assays targeting the same protein may recognize different proteoforms, potentially affecting comparisons between technologies and interpretation of results.

Limitations:

• The atlas is an initial reference, not a complete account of the plasma proteome.
• Most proteoforms were separated by physical properties but not fully identified at the molecular level.
• Low-abundance and disease-associated forms may be underrepresented, and some results may represent groups of related proteoforms rather than individual molecules.
• Some detected proteins may have resulted from cellular contamination.

Conclusion:

The authors state that further work using top-down mass spectrometry, clinical cohorts, and samples from different populations is needed to determine which proteoforms relate to normal biology or disease.

Sources:

• Article context describing a bioRxiv preprint