Mnova BioHOS (User Manual)

Significance of the Topic

High-order structure (HOS) characterization of monoclonal antibodies and recombinant proteins is critical for ensuring safety, efficacy, and batch-to-batch consistency in biopharmaceutical production. Among analytical techniques, nuclear magnetic resonance (NMR) offers atomic-level sensitivity to conformational changes and has become a benchmark for HOS assessment. Robust software tools that integrate spectral processing, comparison, and chemometric analysis accelerate interpretation and support regulatory quality control.

Objectives and Study Overview

This manual describes the functionality and workflow of Mnova BioHOS 3.1, a dedicated NMR chemometrics plug-in for HOS evaluation. The document reviews installation, licensing, data import, processing routines, and four main comparison methodologies: ECHOS, CCSD, 1D Profile, and multivariate statistics (PCA, SIMCA, PLS). Key goals are to provide reproducible pipelines for spectral stacking, region selection, automated alignment, and statistical modeling to differentiate reference and test samples.

Methodology and Instrumentation

Data are acquired from high-field NMR spectrometers (e.g., Bruker, Agilent/Varian, JEOL) using 1D or 2D pulse sequences (e.g., 1H-13C HSQC). Mnova software automates zero-filling, apodization, phasing, baseline correction, ridge removal, and noise reduction. Stacked spectrum tables facilitate class assignments and color coding for sample groups. Region-of-interest definitions (blind regions and cuts) and reference alignment ensure consistent binning. Chemometric modules implement:

• ECHOS: intensity scatter plots and correlation coefficients for pairwise spectrum comparison.
• CCSD: peak-pair chemical shift deviation and intensity ratio analysis based on 2D peak picking.
• 1D Profile: fingerprint subtraction approach with spectral broadening and statistical cross-correlation of repeated measurements represented by box-whisker plots.
• Multivariate Statistics (PCA, SIMCA, PLS): data reduction, classification, regression modeling, and distance metrics for batch discrimination.

Main Results and Discussion

Each analysis method produces quantitative metrics of sample similarity:

• ECHOS yields correlation coefficients (R) and residual heat maps highlighting local spectral differences.
• CCSD reports average deviation values (ppb), amplitude ratios, and combined plots to identify shifting and broadening at specific cross-peaks.
• 1D Profile generates matching factors for intra- and inter-class comparisons, flagging outlier production batches.
• PCA models visualize sample clusters, influence/outlier plots, loading spectra correlations, and Mahalanobis distances; SIMCA defines class boundaries via Cooman’s plots; PLS predicts response variables with R2/Q2 diagnostics and VIP scores pinpointing key spectral regions.

Benefits and Practical Applications

The integrated Mnova BioHOS workflow streamlines complex NMR data handling, reduces manual intervention, and provides transparent statistical validation. It supports:

• Quality control release testing of biotherapeutics.
• Comparability studies for biosimilars and stability monitoring.
• Regulatory submissions by delivering traceable metrics and visual reports.

Future Trends and Potential Applications

Advances may include machine-learning models trained on historical HOS datasets to predict conformational risks, deeper integration with online process analytical technology (PAT) platforms, and expansion to other spectroscopic modalities (e.g., mass spectrometry). Automated adaptive region selection and cloud-based collaborative analytics promise further acceleration of HOS assessment.

Conclusion

Mnova BioHOS 3.1 combines robust NMR processing with versatile chemometric tools to deliver comprehensive HOS evaluation. Its modular approach accommodates diverse experimental designs and ensures reproducible, statistically sound comparisons essential for modern biopharmaceutical development and quality assurance.

References

• Kiss R., Fizil Á., Szántay C. Journal of Pharmaceutical and Biomedical Analysis, 2018, 147:367–377.
• Brinson R.G. et al. mAbs, 2018, 11(1):94–105.
• NIST Monoclonal Antibody Reference Material 8671, National Institute of Standards and Technology.
• Bruker Pharma Journal Club, Bruker BioSpin.