Evaluation of New Peak Detection Performance in MAM Using a Controlled Peptide Spike-in Strategy

Significance of the topic

The Multi-Attribute Method (MAM) augmented with New Peak Detection (NPD) is increasingly used for comprehensive quality monitoring of biotherapeutics. NPD provides untargeted surveillance for unexpected variants or impurities that are not covered by predefined critical quality attribute (CQA) assays. Robust evaluation of NPD performance is essential for regulated environments because sensitivity and specificity must be balanced to avoid missed low-level impurities or excessive false positives from noise and matrix artifacts.

Study objectives and overview

This study introduces a controlled peptide spike-in strategy to systematically characterize NPD performance within a MAM software environment. The goals were to quantify sensitivity and specificity across spike levels, evaluate the impact of detection (LOD) and fold-change thresholds on true- and false-positive reporting, and demonstrate the practical use of integrated NPD reporting for routine review and decision-making.

Methodology

• Reference matrix: a tryptic digest of NIST monoclonal antibody (mAb) at 0.5 µg on column.
• Spike-in standard: PRTC peptide mixture added to the NIST mAb digest at levels of 0.5–4 pmol per injection (≈0.2–1.5% of total digest), enabling controlled low-level impurity simulation.
• Data acquisition: peptide mapping LC/MS in positive ion mode with data-dependent high-resolution MS1 detection across m/z 200–3000 and 1 spectrum per second acquisition rate.
• Data processing: MS1 features extracted and refined by charge, isotope pattern, LC linking, and adduct detection. An internally developed NPD algorithm in MAM software performed sample-to-reference matching using mass accuracy and retention time tolerances, applied intensity (LOD) and fold-change thresholds, and produced structured reports for manual review.
• Performance metrics: sensitivity (ability to detect spiked peptides at low levels), specificity (absence of spurious flagged features when comparing replicates), and fold-change accuracy (quantitative proportionality between spike levels).

Used instrumentation

• Mass spectrometer: Agilent 6230C Time-of-Flight (TOF) instrument equipped with an Agilent Dual AJS electrospray ionization source operated in positive polarity.
• LC column and conditions: AdvanceBio Peptide Mapping C18 column (2.1 × 150 mm) with a water/ACN gradient containing 0.1% formic acid; 0.4 mL/min flow, 40 °C column temperature, and an 84-minute method optimized for peptide mapping separations.
• Key MS settings summarized: MS1 m/z range 200–3000, acquisition 1 Hz, fragmentor ~45 V, typical sheath/drying gas flows and temps for robust ESI operation.

Main results and discussion

• LOD threshold effects: A lower intensity threshold improved recovery of low-level spiked peptides. The study reported consistent detection of 15 PRTC peptides at an LOD around 4,500 intensity units, while a more stringent threshold (e.g., ~11,000) led to missed detections at the lowest spike levels (0.5 pmol).
• Specificity: Replicate injections of the reference matrix compared against the reference produced zero flagged features (no New, Increased, Decreased, or Missing calls) under optimized NPD settings, demonstrating high specificity and low false-positive reporting in controlled conditions.
• Fold-change accuracy: Comparison of 1 pmol versus 2 pmol spikes showed proportional increases in extracted ion chromatogram (EIC) peak areas and consistent fold-change values close to the expected ~2×, indicating that the NPD workflow preserved reliable relative quantitation for trend analysis.
• Review interface: The MAM software provided integrated visualization (sample overview, feature tables, EIC and MS spectral views) and structured reporting to support verification of flagged features and streamline pass/fail decisions.

Benefits and practical applications

• The controlled spike-in strategy offers a standardized way to benchmark NPD sensitivity and specificity, informing threshold selection and tuning for regulated workflows.
• Optimized LOD and fold-change settings reduce the risk of false positives while enabling detection of relevant low-level impurities, improving confidence in product purity monitoring.
• Integrated reporting and visualization expedite expert review and support routine release or stability testing where both targeted CQA monitoring and untargeted surveillance are required.

Future trends and potential applications

• Advances in machine learning and improved feature-filtering logic could further reduce false positive rates by distinguishing chemical noise and adducts from genuine low-abundance variants.
• Standardized spike-in panels and inter-laboratory studies could establish community benchmarks for NPD performance and regulatory acceptance.
• Higher sensitivity instrumentation and improved chromatographic resolution will enable detection of even lower-level impurities while preserving specificity when coupled with smarter data-processing workflows.
• Extension of controlled spike-in strategies to other molecule classes (e.g., intact proteins, glycoforms, aggregates) would broaden applicability of NPD validation across biopharma analytics.

Conclusion

The presented controlled peptide spike-in approach provides a practical framework to characterize NPD performance within a MAM environment. By systematically varying spike levels and tuning intensity and fold-change criteria, the study demonstrated a workable balance between sensitivity and specificity, confirmed accurate fold-change quantitation, and showed how integrated reporting supports routine decision-making. These findings support adoption of rigorously validated NPD workflows in regulated biotherapeutic analytics while highlighting areas for continued method development and standardization.

References

• Rogers RS et al., Anal. Chem., 2015, 87, 1058–1065.
• Ren D et al., mAbs, 2018, 10, 859–870.
• USP <1060> and ICH Q2(R2) guidance on analytical method validation and system suitability considerations.