Charge Assignment of Isotopically Resolved Direct Mass Data by 2D Voting

Importance of the Topic

Accurate molecular weight determination of proteins and complex biomolecules relies on precise charge assignment of individual ions in mass spectrometry. Variability in charge estimation can obscure isotopic resolution and hinder identification of minor species. The development of robust computational methods to assign unique charges enhances the analytical power of direct single-ion data acquisition modes and supports high-resolution proteoform analysis.

Objectives and Overview of the Study

This study introduces a two-dimensional voting algorithm to refine charge assignments in Direct Mass Technology Mode. The method connects ions related by isotopic spacing and charge shifts in the m/z–Z space, iteratively removing incorrect estimates and converging on unique charge states. The algorithm was applied to three datasets: native and denatured SILu™Lite Universal Antibody Standard and native Enbrel® (Etanercept).

Methodology

Sample Preparation and Data Acquisition

• SILu and Enbrel were buffer-exchanged into 100 mM ammonium acetate using centrifugal filters and diluted to final concentrations of 30 µM and 200 nM, respectively.
• Direct single-ion spectra were collected on a Thermo Q Exactive UHMR with STORI slope tracking to estimate charge.

2D Voting Algorithm
The algorithm maps individual ions into bins defined by m/z and preliminary integer charge Z. For each bin, trial charge states within ±5% are considered. Votes are cast from neighboring bins (m/z offsets corresponding to isotopic mass differences and charge shifts) weighted by bin occupancy and probability. After tallying votes, probabilities are normalized and the top three charge candidates are retained. This process iterates 6–9 times, and the highest-probability assignment above a threshold (0.5) is selected as the final charge.

Instrumentation Used

• Mass Spectrometer: Thermo Scientific Q Exactive UHMR.
• Ionization: Static nano-ESI tip at 1.1–1.4 kV, capillary at 300 °C.
• Acquisition Settings: RF level 150%, in-source trapping off, HCD pressure 0.1–0.5.
• Data Processing: STORI files collected via Tune/UHMR; voting algorithm implemented in Python within STORIboard.

Main Results and Discussion

After eight iterations, unique charge assignments were achieved for all three datasets. Native SILu data (≈2.5 million ions) revealed over 30 overlapping charge states with clear isotopic separation of glycoforms. Denatured SILu (≈500 000 ions) showed reduced complexity and resolved minor components (C1, C2, C3, dimers, double ions). Native Enbrel (≈100 000 ions) produced an isotopically resolved molecular weight distribution spanning 127 k–135 k Da with at least seven glycoforms separated by hexose increments.

Benefits and Practical Applications

• Enhanced spectral clarity by eliminating charge ambiguity and reducing coefficient of variation.
• Detection of low-abundance proteoforms, dimers, and modified species in complex mixtures.
• Compatibility with high-throughput workflows for biopharmaceutical characterization and QA/QC.

Future Trends and Potential Applications

Integration of the voting algorithm into real-time acquisition software could enable on-the-fly charge assignment and immediate feedback during experiments. Expanding the approach to more complex proteomes and coupling with advanced machine learning may further improve speed and accuracy. Adaptation to other single-ion detection platforms could broaden its applicability in native and denatured mass spectrometry.

Conclusion

The 2D voting method provides a systematic, data-driven strategy to assign unique charges to isotopically resolved single-ion mass spectra. By leveraging the precise relationships among isotopic and charge neighbors, the algorithm yields high-resolution molecular weight distributions and reveals minor components in complex samples with minimal manual intervention.

References

1. Kafader JO, Beu SC, Early BP, et al. Journal of the American Society for Mass Spectrometry 2019, 30, 2200–2203.
2. Schachner LF, Ives AN, McGee JP, et al. Journal of the American Society for Mass Spectrometry 2019, 30, 1190–1198.