Aggregation is one of the main reasons of false positives in HTS. To build a reliable prediction model, we collected a large dataset of aggregators and drugs to build several models based on the combination of different descriptors and algorithm, ChemAGG is a free webserver that could be used to easily filter out aggregators from potential lead molecules!
Model performance: Training set: 31,316; Test set: 4,975; Descriptor: ECFP4; Accuracy(CV): 0.925; AUC(CV): 0.970; Specificity(CV): 0.960; Sensitivity(CV): 0.853; Accuracy(test): 0.922; AUC(test): 0.968; Specificity(test): 0.963; Sensitivity(test): 0.822;
| By Inputting SMILES
| By Uploading Files (*.sdf) Format converter
| By Drawing Molecule from Editor Below
| Select the Data Source