ADME
Pharmaco-kinetics
Task Overview
ADME (Absorption, Distribution, Metabolism, and Excretion) properties are crucial pharmacokinetic parameters that determine how a drug moves through the body. This task aims to predict various ADME properties of drug molecules to assess their suitability as therapeutic candidates.
Impact
Poor ADME properties are a major cause of drug failure in clinical trials. Early and accurate ADME prediction can significantly reduce attrition rates and development costs by identifying problematic compounds before expensive clinical testing. This enables pharmaceutical companies to prioritize compounds with favorable pharmacokinetic profiles.
Generalization
As drug structures evolve over time, ADME prediction models must generalize to novel chemical scaffolds with limited structural similarity to existing training data. Models need to capture the complex relationships between molecular structure and pharmacokinetic behavior.
Product
Small-molecule.
Pipeline Stage
Lead development and optimization.
Available Datasets
Caco2_Wang
PAMPA_NCATS
HIA_Hou
Pgp_Broccatelli
Bioavailability_Ma
Lipophilicity_AstraZeneca
Solubility_AqSolDB
HydrationFreeEnergy_FreeSolv
BBB_Martins
PPBR_AZ
VDss_Lombardo
CYP2C19_Veith
CYP2D6_Veith
CYP3A4_Veith
CYP1A2_Veith
CYP2C9_Veith
CYP2C9_Substrate_CarbonMangels
CYP2D6_Substrate_CarbonMangels
CYP3A4_Substrate_CarbonMangels
Half_Life_Obach
Clearance_Hepatocyte_AZ
Usage Example
You can access these datasets using the PyTDC library:
from tdc_ml.single_pred import ADME
# Load a dataset
data = ADME(name='Caco2_Wang')
# Access the data
df = data.get_data()
print(df.head())