Bilinear decomposition based alignment of chromatographic profiles.

Area of Science:

• Analytical Chemistry
• Chemometrics
• Chromatography

Background:

• Chromatographic signal alignment is crucial for accurate data analysis.
• Existing alignment methods can be prone to errors when dealing with complex signals.
• Photodiode array detection generates multi-way data requiring sophisticated processing.

Purpose of the Study:

• To develop a novel, robust alignment procedure for chromatographic signals.
• To improve the accuracy of chromatographic data analysis by minimizing alignment errors.
• To effectively handle complex chromatographic data from photodiode array detection.

Main Methods:

• Utilized multivariate curve resolution-alternating least squares (MCR-ALS) for chemometric resolution.
• Applied a spectral selectivity constraint to differentiate background and constituent spectra.
• Employed correlation optimized warping (COW) for individual alignment of resolved chromatographic profiles.
• Reconstituted original data with aligned profiles and corresponding spectra.

Main Results:

• Successfully decomposed multi-way data into a bilinear model of chromatographic profiles and spectral signatures.
• Obtained pure concentration profiles through MCR-ALS with effective background differentiation.
• Achieved improved alignment accuracy compared to standard correlation optimized warping.
• Demonstrated the procedure's effectiveness on two sets of chromatographic signals.

Conclusions:

• The novel alignment procedure offers enhanced accuracy for chromatographic signal analysis.
• Integrating MCR-ALS with COW minimizes risks associated with aligning noncorresponding chromatographic information.
• This method provides a more reliable approach for processing complex chromatographic data from photodiode array detection.