How AI and Machine Learning Are Reshaping Separations Science at HPLC 2026

On Wednesday, June 10, a session on Artificial Intelligence and Machine Learning in Separations will be held at HPLC 2026 from 11:00am–12:30pm in the White River E room at the JW Marriott Indianapolis in Indianapolis, Indianapolis.

The session will be chaired by Robert Kennedy from the University of Michigan.

A keynote presentation, AI-Augmented Liquid Chromatography-Mass Spectrometry in Metabolomics: Promoting Methodological Innovation and Clinical Translation, from Guowang Xu of the Dalian Institute of Chemical Physics, Chinese Academy of Sciences (China), will be presented from 11:00am–11:30am. This talk will examine how the integration of AI with advanced LC–MS platforms is transforming metabolomics across the entire analytical workflow, from data acquisition and feature extraction to metabolite annotation and biological interpretation. Addressing key limitations of conventional untargeted LC–MS approaches, such as slow data processing, batch effects, and low structural identification rates, the presentation will highlight how AI-enabled tools improve throughput, reproducibility, data pretreatment, and dark matter annotation efficiency. Central to this work is the development of a standardized, automated, and intelligent "third-generation" metabolomics framework, built on robust LC–MS methodologies, enabling large-scale, high-coverage metabolic profiling. The talk also explores the clinical applications of these advances through collaborative research on type 2 diabetes, obesity, and related metabolic diseases, demonstrating how the combination of LC–MS data with multi-omics inputs can reveal novel biomarkers, uncover metabolic reprogramming patterns, and inform potential interventions including dietary modulation, pharmacological therapies, and surgical approaches.

Following this, Imad Haidar Ahmad of Amgen (USA) will present his keynote on Retention Time-Aligned Method Development Through In Silico Chromatography Modeling from 11:30am–11:50am. This presentation will address the growing analytical demands placed on pharmaceutical laboratories by new drug substances, particularly biopharmaceuticals and multi-combination drugs, where the complexity of analysis extends well beyond manufacturing challenges. He will showcase novel tools for streamlining chromatographic method development, beginning with systematic exploration of column and mobile phase combinations to identify optimal conditions, followed by in silico simulation using a minimal number of experiments to maximize resolution between target peaks. At the heart of this approach is a computer-assisted modeling framework that enables the development of retention time-aligned chromatographic methods, supporting robust and efficient method transfer across different experimental conditions, chromatography systems, and laboratory settings.

From 11:50am–12:10pm, An Ensemble Machine Learning Approach for Custom Chromatographic Peak Integrationwill be presented by Scott Trinkle of Waters Corporation (USA).

The last presentation, From Experiments to Algorithms: AI-Powered Digital Transformation of Chromatographic Method Development, will be led by Pankaj Aggarwal of Merck (USA), and will be held from 12:10pm–12:30pm. This talk will address the longstanding empirical and resource-intensive nature of chromatographic method development across pharmaceutical discovery and manufacturing, making the case for a fundamental shift toward AI-enabled, end-to-end digital workflows. Drawing on industrial case studies spanning analytical and preparative scales across multiple development stages, the presentation will demonstrate how the integration of physicochemical understanding with machine learning, Bayesian optimization, and predictive analytics can transition chromatographic workflows from sequential experimentation to data-driven, decision-focused processes. These integrated models act as a digital twin of the LC system, drawing on historical and real-time data to guide both experimental and operational decisions across method screening, optimization, and life cycle monitoring. The result is a reimagining of chromatographic method development as a continuous, self-improving life cycle that bridges experimentation, computation, and industrial execution to deliver more reliable and scalable LC workflows.