ADMET Predictions

ADMET stands for Absorption, Distribution, Metabolism, Excretion, and Toxicity. ADMET predictions refer to the computational estimation of these pharmacokinetic and toxicological properties of compounds. These predictions are crucial for evaluating the drug-likeness of molecules and their potential success as therapeutic agents. If a molecule is potent and selective, the inability to reach its desired target at doses that are not toxic will hamper its progression. The key aspects of ADMET include:

• Absorption: How well a drug is absorbed into the bloodstream.
• Distribution: The distribution of the drug throughout the body and its ability to reach the target site.
• Metabolism: The biochemical modification of the drug, primarily by liver enzymes.
• Excretion: The elimination of the drug from the body.
• Toxicity: The potential harmful effects of the drug or its metabolites on the body.

Importance in Computational Drug Discovery

• Early Assessment: ADMET predictions allow for the early identification of potential issues with a compound’s pharmacokinetics and toxicity, reducing the risk of late-stage failure.
• Cost and Time Efficiency: By predicting ADMET properties computationally, resources can be focused on the most promising compounds, saving time and reducing costs associated with experimental testing.
• Lead Optimization: Helps in the optimization of lead compounds by predicting modifications that can improve ADMET properties.
• Safety Profiling: Ensures that compounds with undesirable toxicological profiles are identified and excluded early in the drug discovery process.
• Regulatory Compliance: Provides data that can support regulatory submissions and the development of safer and more effective drugs.

Key Tools

1. DeepOrigin’s ADMET Prediction Tool: An integrated tool for predicting ADMET properties optimized for drug discovery applications. Available in Balto or via API.

2. ADMET Predictor: A software tool for predicting ADMET properties using machine learning models and molecular descriptors.

3. pkCSM: A tool for predicting pharmacokinetic and toxicity properties of small molecules using graph-based signatures.

4. ADMETlab: An online platform for comprehensive ADMET property prediction.

Literature

Key Papers on ADMET Predictions in Drug Discovery

1. “ADMET in silico modeling: towards prediction paradise?”

• Publication Date: 2010

• DOI:10.1038/nrd1032

• Summary: Reviews the state-of-the-art in ADMET modeling, discussing current methodologies and future prospects.

2. “ADMET prediction in drug discovery: 20 years of progress”

• Publication Date: 2016

• DOI:10.1186/s13321-020-00421-y

• Summary: Reviews the progress in ADMET prediction over the past two decades, highlighting key advancements and challenges.

3. “Hybrid Fragment-SMILES Tokenization for ADMET Prediction in Drug Discovery”

• Publication Date: 2024-08-01

• DOI:10.1186/s12859-024-05861-z

• Summary: Reveals that hybrid tokenization with high-frequency fragments enhances ADMET prediction results.

4. “ADMET Evaluation in Drug Discovery: Application and Industrial Validation of Machine Learning Algorithms for Caco-2 Permeability Prediction”

• Publication Date: 2025-01-10

• DOI:10.1186/s13321-025-00947-z

• Summary: Evaluates machine learning algorithms for predicting Caco-2 permeability, showing that XGBoost provides better predictions for test sets.

5. “ADMET Evaluation in Drug Discovery. 19. Reliable Prediction of Human Cytochrome P450 Inhibition Using Artificial Intelligence Approaches”

• Publication Date: 2019-10-23

• DOI:10.1021/acs.jcim.9b00801

• Summary: Develops classification models using ensemble learning and deep learning to predict CYP450 inhibition, showing that ensemble learning models generally give better predictions.

6. “Evaluation of ADMET Predictor in Early Discovery Drug Metabolism and Pharmacokinetics Project Work”

• Publication Date: 2021-11-08

• DOI:10.1124/dmd.121.000552

• Summary: Evaluates the prediction models for lipophilicity, solubility, metabolic stability, and permeability using ADMET Predictor software.

7. “Overview of Computational Toxicology Methods Applied in Drug and Green Chemical Discovery”

• Publication Date: 2024-12-01

• DOI:10.3390/jox14040101

• Summary: Introduces the development of computational toxicology, focusing on machine learning and deep learning methods in toxicology.