Researchers Predicted Molecule Retention Times via Machine Learning
New 2-step software simplifies chemical analysis by calculating retention orders before mapping to specific times.
Updated on Oct. 5, 2026 in Chemistry
Live Poll
Do you believe new scientific analysis tools make healthcare and environmental testing more trustworthy?
Researchers from Friedrich Schiller University Jena, Helmholtz Zentrum München, and Technical University of Munich have released a new machine learning method called 2-step. The published research details a tool that predicts molecule retention times in liquid chromatography by calculating retention indices.
Why it matters
Accurate retention time prediction is difficult because results depend on variable experimental conditions like pH, temperature, and column type. By automating this process, the method could be integrated into analytical software to streamline laboratory workflows.
The 2-step tool functions by determining a retention order index for a molecule, which is then mapped to precise retention times using known reference points. This approach accounts for specific conditions such as solvent, gradient, and column type to improve prediction reliability.
The players
Friedrich Schiller University Jena
A research university based in Germany that focuses on interdisciplinary scientific studies and advanced chemical analytics.
Helmholtz Zentrum München
A German research institution specialized in health and environmental sciences with a focus on data-driven discovery.
Technical University of Munich
A public research university in Germany known for its strong emphasis on technology and engineering applications.
The details
Liquid chromatography is a technique used to separate and identify components in a mixture based on their interaction with a column. The 2-step machine learning model calculates the predicted sequence in which molecules will exit the column, known as the retention order, before converting these values into specific retention times. The researchers have made this tool available as both a software package and a web application to facilitate broader integration into analytical laboratory programmes.
Timeline
October 2026: The research team published their method in the journal Nature Methods.
The Tech Race
This development follows an ongoing shift toward automated chemical identification workflows that reduce the need for manual calibration. The 2-step tool represents a move to standardize prediction models that were previously dependent on highly specific, lab-tailored conditions.
Laboratory researchers and analytical chemists will gain access to a new web-based tool and software package for predicting retention times. This methodology is designed to be integrated into existing analytical pipelines to automate and improve the speed of chemical analysis workflows.
The takeaway
The study demonstrates that machine learning can reliably bridge the gap between abstract retention indices and actual chromatographic results. Watch for further adoption of the 2-step package in high-throughput labs to see if it reduces the need for frequent instrument recalibration.
Further reading
For broader trends in molecular modeling, explore our recent coverage in Chemistry.
Source note: This article includes information reported by Analytik.
Live Poll
Do you believe new scientific analysis tools make healthcare and environmental testing more trustworthy?
