Multivariate analytics of chromatographic data: Visual computing based on moving window factor models.

Area of Science:

• Biotechnology
• Analytical Chemistry
• Data Science

Background:

• Chromatography is crucial in biotechnology, generating vast data from new devices driven by Process Analytical Technology (PAT) and Quality by Design (QbD).
• Analyzing large chromatographic datasets is challenging, often leaving valuable information undiscovered within big data.

Purpose of the Study:

• To present a novel, systematic, top-down approach for analyzing complex chromatographic datasets.
• To develop a method that highlights outliers, drifts, and data inconsistencies within multi-dimensional chromatographic data.
• To facilitate the localization of these anomalies in a straightforward manner.

Main Methods:

• Utilized a top-down strategy focusing on global information first.
• Employed moving window factor models to extract key differences between samples.
• Developed an interactive visualization tool for explorative analysis of complex datasets.

Main Results:

• The approach effectively highlights interesting regions like outliers and drifts in chromatographic data.
• The interactive tool simplifies localizing variances within multi-dimensional datasets.
• The method successfully differentiates between explainable and unexplainable variance in the data.

Conclusions:

• The developed workflow provides a starting point for detailed raw data analysis by generating temporally dependent difference descriptors.
• This approach enhances the systematic analysis of chromatographic big data in biotechnology.
• The tool offers a convenient way to explore and understand complex chromatographic datasets.