Chromatographic peak alignment using derivative dynamic time warping.

Area of Science:

• Bioprocess engineering
• Analytical chemistry
• Data science

Background:

• Chromatogram overlays are vital for monitoring bioprocess purification but are challenging to analyze objectively due to peak shifts.
• Variable phase durations and process holds in biomanufacturing introduce variability, complicating direct comparison of chromatograms.
• Synchronization of batch process data is often necessary before applying advanced analytical techniques like multivariate analysis.

Purpose of the Study:

• To introduce dynamic time warping (DTW) with a derivative weighting function for objective chromatogram analysis.
• To demonstrate the utility of DTW in aligning time-variant chromatographic signals from bioprocesses.
• To showcase DTW as an effective preprocessing step for multivariate model development in biomanufacturing.

Main Methods:

• Application of dynamic time warping (DTW), a technique originally for speech recognition, to align chromatograms.
• Utilizing a derivative weighting function within DTW to enhance the alignment of chromatographic peaks.
• Implementing DTW as a preprocessing step for multivariate data analysis.

Main Results:

• DTW successfully aligned chromatograms despite peak shifts caused by process variations.
• The derivative weighting function improved the accuracy of chromatogram alignment.
• DTW facilitated objective process monitoring and fault detection by enabling direct comparison of aligned chromatograms.

Conclusions:

• Dynamic time warping provides an objective method for analyzing chromatograms in bioprocess monitoring.
• This technique effectively addresses challenges posed by peak shifts and variable process durations.
• DTW serves as a valuable preprocessing tool for developing robust multivariate models for biomanufacturing data analysis.