Fig. 6: Definition and application of Chromatographic enantioseparation probability assessment.
a The definition of separation probability and some examples under different conditions, including an inseparable situation under the ADH column and two separate situations under ODH and OJH columns. The red cross refers to the observed data while the lines in error bar refers to 90th quantile, predicted retention time (RT), and 10th quantile, respectively, which are predicted by the model. b The violin plot of the distribution of calculated Sp for 412 pairs of testing enantiomers. The black lines mean the maximum, median, and minimum values, respectively. c The accuracy of enantioseparation prediction for the enantiomers with different thresholds of similarities. d An example of the utilization in practical application, including prediction of retention time and Sp for the same enantiomers under different columns made by multi-column prediction model (left), and verification experiments (right). Sp refers to the separation probability. e The \(\triangle {RT}\) for each of the candidate conditions with four conventional ML methods. The dark dotted line refers to the separation threshold. Source data are provided as a Source Data file.
