Organizers
Powered by
| Introduction of DL-AOT |
DL-AOT prediction server is a free web tool to evaluate acute oral toxicity (AOT) of small molecules. It is based on the consensus regression model (deepAOT-R) and the best multi-classification model (deepAOT-C) constructed by a deep learning framework, which is composed of multi-layer convolution neural networks. Performances were measured by two external test sets contains 1673 and 375 compounds, separately. The R2 and MAE of deepAOT-R was 0.864 and 0.195 for test set I, and the accuracy of deepAOT-C was 95.5% and 96.3% for the test sets I and II, respectively, which showed the state-of-the-art level for AOT.
For details about development and validation of the two models, please refer to this manuscript: Deep Learning Based Regression and Multi-class Models for Acute Oral Toxicity Prediction with Automatic Chemical Feature Extraction, 2017.
Developed and maintained by the Center of Quantitative Biology (CQB) and Molecular Design Laboratory (MDL) of PKU | Peking University.
For details about development and validation of the two models, please refer to this manuscript: Deep Learning Based Regression and Multi-class Models for Acute Oral Toxicity Prediction with Automatic Chemical Feature Extraction, 2017.
Developed and maintained by the Center of Quantitative Biology (CQB) and Molecular Design Laboratory (MDL) of PKU | Peking University.
*Note.
1. The format of an input file is allowed, such as "*.smi", "*.mol", "*.mol2", "*.sdf"; The size of an input file is less than 200KB;
2. The "Regression model and multi-class model" option from "Model Type": use both of regression consensus model and multi-classification model predict the input small molecules.
3. This web server is freely accessible to academics. For industry, please send inquires to jfpei@pku.edu.cn.