Times are changing but order matters: transferable prediction of small-molecule liquid chromatography retention times.

A two-step machine learning approach enables transferable prediction of small-molecule retention times across different chromatographic conditions, outperforming existing methods.

• Why it matters: Accurate retention time prediction is crucial for small-molecule analysis, but current models struggle to generalize across varying chromatographic systems, limiting their practical utility.
• What they did: The authors developed a two-step method: first predicting retention order indices considering chromatographic conditions, then mapping these indices to absolute retention times, without needing target-system training data.
• The result: This approach achieves superior transferability and accuracy compared to existing models, and provides insights into how chromatographic conditions influence retention order changes.