Pretrained Models

BBB

• BBB-score:
  
  
  

CYP450

• D-CyPre: Tool for Accurately Predicting the Site of Metabolism of Human Cytochrome P450 Metabolism
  
  
  

General

• ADMET_XGBoost:
  
  
  
• admet-ai:
  
  
  
• ADMETlab:
  
  
• ADMETlab3.0:
  
• ChemMORT:
  
  
  
• Computation-ADME:
  
  
• QIP: Machine learning model that predict the ADMET (absorption, distribution, metabolism, excretion, and toxicity) properties of molecules.
  
  
  

Metabolites

• MetaPredictor: in silico prediction of drug metabolites based on deep language models
  
  
  

PK

• PKSmart: This work used molecular structural fingerprints, physicochemical properties, and predicted animal PK data as features to model the human PK parameters VDss, CL, t½, fu and MRT for 1,283 unique compounds and developed a webhosted application PKSmart, the first work that publicly releases PK models on par with industry-standard models.
  
  

Toxicity

• ProTox3: Webserver for toxicity prediction
  
  

hERG

• CardioTox: A robust predictor for hERG channel blockade via deep learning meta ensembling approaches
  
  
• CLOP-hERG: Contrastive Learning Optimized Pretrained Model for Representation Learning in Predicting Drug-Induced hERG Channel Blockers
  
  
• hERGdb: web-based cardiotoxicity prediction
  

logP

• rescoss_logp_ml: A repository dedicated to logP prediction using machine learning, offering insights into the solubility and distribution properties of compounds.
  
  
  .)