AICpred

AICpred is a machine learning-based web application for predicting small molecules as potential inhibitors against the binding of TLR4-MyD88. Given the major role this binding plays in pro-inflammatory mechanisms, the predicted compounds can be explored experimentally as possible therapeutic agents for ameliorating cytokine storm. The models were trained on 195,980 experimentally-tested TLR4 inhibitors and non-inhibitors. The web application offers predictions using the XGBoost model, which achieved a balanced accuracy of 0.958 and AUROC score of 0.999.

Predict