Prediction and Preparation of Co-amorphous Phases of a Bislactam [dataset]

The effectiveness of a partial least squares – discriminant analysis co-amorphous prediction model was tested using a co-amorphous screening data for the dimer of N vinyl(caprolactam) (bisVCap) as co-former with a range of active pharmaceutical ingredients. The prediction model predicted 71% of the systems correctly. An experimental co-amorphous screen was performed with this co-former with 13 different active pharmaceutical ingredients and the results were compared to the prediction model. A total of 85% of the systems correctly predicted. Stability assessments of three co-amorphous systems showed that the prediction model does not provide information on the stability of the co-amorphous material. The model was shown to perform well with small molecules which were different to the training set used to produce it.