RT-Pred

RT-Pred predicts liquid chromatography retention times for chemicals across diverse chromatographic methods using machine learning models, including custom user-trained predictors and a database of millions of predicted retention times spanning over 40 chromatographic methods.

Key Features:

• Custom Model Training: Users can train chromatographic method-specific RT-Pred models using their own retention time data.
• Retention Time Database Search: RT-Pred supports retention time and compound searches against a database of millions of predicted retention times spanning more than 40 different chromatographic methods.
• Void/Retained/Eluted Classification: RT-Pred includes a classifier that accurately identifies compounds eluting in the void volume, retained on the column, or normally eluted, which significantly improves prediction performance.
• In-house Predictors: RT-Pred provides its own chromatographic method-specific retention time predictors for direct use.

Scientific Applications:

• LC-MS Compound Identification: RT-Pred enables retention time prediction from known or postulated chemical structures to improve compound identification in liquid chromatography-mass spectrometry workflows.
• Chromatographic Method Optimization: RT-Pred allows researchers to predict retention times across hundreds of different chromatographic methods with interdependent parameters.
• Void Volume and Column Retention Prediction: RT-Pred identifies compounds that will elute in the void volume or be retained on the column, aiding in separation optimization.

Methodology:

RT-Pred uses machine learning to predict retention times from chemical structures. The method incorporates a void/retained/eluted classifier and achieves prediction through chromatographic method-specific models trained on retention time data aligned with structural data.