Outline

4/19/02


Click here to start


Table of Contents

PPT Slide

Outline

ADME/Tox Property Prediction: Challenges in Medicinal Chemistry

Traditional Drug Discovery Scheme

In silico prediction of ADME properties

QSPR/ADME Modeling Procedure

Chemistry In/Out Modeling

PPT Slide

TAE Descriptors

TAE descriptors continued ...

Electron Density-Derived Descriptors

PPT Slide

WCD Descriptors: Discrete Wavelet Transformation

Discrete Wavelet Transformation: Pyramid Algorithm

Wavelet Representations of High-Resolution Molecular Surface Property Densities. (1,2,10 and 20 Coefficient Decompositions)

Electron Density-Derived Property Modeling: Wavelet Coefficient Descriptors (WCDs)

Electron Density-Derived Molecular Properties: Wavelet Coefficient Descriptors (WCDs)

PPT Slide

PPT Slide

PEST descriptors continued ...

PPT Slide

Data Analysis Techniques

Approaches to feature selection

Practical Issues

Validation Methodology

Validation Methodology continued...

PPT Slide

ADME Problem: Human Intestinal Absorption

Approximate ADME Prediction

Optimization of bioavailability by using the “rule of five”

Bioassay: Caco-2 Permeability

Intestinal Absorption Mechanisms

Human Intestinal Absorption Barriers

The Caco-2 cell line

Caco-2 Assays

Public Caco-2 Data Sources

LSER and TLSER methods

Molecular Polar Surface Area (PSA and PSAd)

Monte Carlo Simulation methods

Quantum Mechanics-Based Descriptors

Atom-based descriptors

Other methods involved in ADME prediction

Rapid Electron Density-Derived Descriptor Methods – DDASSL

Bagged Non-linear SVM (RBF) Models Caco-2 : 718 TAE/PEST Descriptors

Bagged Non-linear SVM (RBF) Models Caco-2 : 31 Variables

PPT Slide

Star Plot of ABSDRN6

Bagged Non-linear SVM (RBF) Models Caco-2 : 15 Variables

Visualization of 15 Selected Descriptors

Analysis: Starplot of a.don

Analysis: Starplot of SlogP.VSA0

Chemical Insights

PPT Slide

Future Directions

PPT Slide

Acknowledgements

That’s all folks…

Wavelet Representations of High-Resolution Molecular Surface Property Densities.

Flowchart of SVMs

PPT Slide

PPT Slide

Hybrid TAE/SHAPE

PPT Slide

Summary

Future Development

PPT Slide

Wrapper Feature Selection Results

Iterative Bagged Feature Selection Results

PPT Slide

Atom-based descriptors

LSER and TLSER methods

Molecular Polar Surface Area (PSA and PSAd)

Monte Carlo Simulation methods

Quantum Mechanics-Based Descriptors

Other methods involved in ADME prediction

Recent Results

Author: Curt M. Breneman

Email: brenec@rpi.edu

Home Page: http://www.drugmining.com