Related Experiment Videos
In silico ADME modelling 2: computational models to predict human serum albumin binding affinity using ant colony
Sitarama B Gunturi1, Ramamurthi Narayanan, Akash Khandelwal
1Life Sciences R&D Division, Advanced Technology Centre, Tata Consultancy Services Limited, # 1, Software Units Layout, Madhapur, Hyderabad 500 081, India.
Bioorganic & Medicinal Chemistry
|March 1, 2006
Summary
Predictive models for human serum albumin (HSA) binding were developed using quantitative structure-property relationship (QSPR) analysis. These models, based on molecular descriptors, accurately predict drug binding affinity, aiding virtual screening.
Area of Science:
- Computational Chemistry
- Medicinal Chemistry
- Pharmacokinetics
Background:
- Human serum albumin (HSA) is a key determinant of drug pharmacokinetics.
- Accurate prediction of drug-HSA binding is crucial for drug development.
- Existing models may lack applicability across diverse chemical spaces.
Purpose of the Study:
- To develop global quantitative structure-property relationship (QSPR) models for in vitro HSA binding.
- To create predictive models applicable to a wide range of medicinal chemistry compounds.
- To identify key molecular descriptors influencing HSA binding affinity.
Main Methods:
- Modelling of HSA binding data for 94 diverse drugs and drug-like compounds.
- Utilized ant colony systems and multiple linear regression (MLR) for descriptor selection.
- Derived QSPR models using a pool of 327 molecular descriptors.
Main Results:
- Developed optimal QSPR models with five and six descriptors exhibiting excellent predictive power.
- Best five-descriptor model: R=0.8942, Q=0.86790.
- Best six-descriptor model: R=0.9128, Q=0.89220, demonstrating high predictive accuracy.
Conclusions:
- HSA binding affinity is primarily governed by hydrophobicity, solubility, size, and shape.
- The developed QSPR models are valuable tools for virtual screening.
- These models facilitate efficient selection and prioritization of drug candidates.