Outlier detection for the Generalized Rank Annihilation Method in HPLC-DAD analysis

Area of Science:

• Analytical Chemistry
• Chromatography
• Spectroscopy

Background:

• Second-order calibration methods like GRAM are crucial for quantifying analytes that coelute with interferences in chromatography.
• Accurate quantification relies on analyte peaks in standards and test samples having identical shapes and retention times (trilinear structure).
• Deviations in peak shape or retention time can lead to outliers and incorrect predictions, which are not always detectable by standard GRAM output checks.

Purpose of the Study:

• To introduce novel graphical methods for detecting outliers in GRAM-based chromatographic quantification.
• To address the limitations of existing GRAM checks in identifying peak shape and retention time variations.
• To improve the reliability and accuracy of analyte quantification when dealing with complex sample matrices.

Main Methods:

• Development and application of graphical plots to compare elution profiles recovered by GRAM.
• Utilizing GRAM-derived elution profiles and spectra to define interference vector spaces.
• Projecting measured peaks onto orthogonal spaces and checking for proportionality using singular vectors or orthogonal signal versus net sensitivity plots.

Main Results:

• Demonstrated the effectiveness of the proposed plots in identifying outliers caused by variations in peak alignment and shape.
• Validated the graphical methods using simulated data, confirming their ability to detect deviations from trilinearity.
• Successfully applied the methods to quantify 4-nitrophenol in river water samples using liquid chromatography/UV-Vis detection, showcasing real-world applicability.

Conclusions:

• The presented graphical methods provide a reliable means to detect outliers in GRAM analyses that arise from peak shape and retention time mismatches.
• These diagnostic tools enhance the robustness of GRAM for accurate analyte quantification in chromatography, especially in the presence of interferences.
• The study offers practical solutions for improving data quality and interpretation in complex chromatographic analyses.