Computational Methods for Predicting ADMET Properties: Tools and Strategies to Minimize Late-stage Drug Failures
Computer-Aided Drug Design: Principles, Techniques, and Applications, wiley, ss.187-198, 2026
- Yayın Türü: Kitapta Bölüm / Araştırma Kitabı
- Basım Tarihi: 2026
- Doi Numarası: 10.1002/9781394371112.ch09
- Yayınevi: wiley
- Sayfa Sayıları: ss.187-198
- Anahtar Kelimeler: ADMET, In silico drug discovery, Pharmacokinetics, Predictive toxicology, SwissADME, Virtual screening
- Atatürk Üniversitesi Adresli: Evet
Özet
The high attrition rate in drug development, particularly during clinical phases, is predominantly attributed to unfavorable absorption, distribution, metabolism, excretion, and toxicity (ADMET) profiles. Approximately 40–50% of drug candidates fail due to poor pharmacokinetic properties and toxicity issues. This chapter examines three key in silico platforms, pkCSM, SwissADME, and the Toxicity Estimation Software Tool (TEST), that enable early prediction of ADMET characteristics, thereby minimizing late-stage failures. These computational tools facilitate the prioritization of compounds with optimal bioavailability and safety profiles before costly experimental validation. Through comparative analysis and case studies, this chapter demonstrates how the integrated use of these platforms can streamline drug discovery, reduce development costs, and accelerate the delivery of safer therapeutics to market. The discussion also addresses current limitations and future directions in computational ADMET prediction, including the integration of artificial intelligence and machine learning approaches.