Related Experiment Videos
Prediction of Human Intestinal Absorption of Compounds Using Artificial Intelligence Techniques
Rajnish Kumar1, Anju Sharma1, Mohammed Haris Siddiqui2
1Amity Institute of Biotechnology, Amity University Uttar Pradesh, Lucknow, 226028, Uttar Pradesh. India.
Current Drug Discovery Technologies
|April 7, 2017
Summary
Support vector machine (SVM) models offer superior accuracy in predicting human intestinal absorption compared to other artificial intelligence methods. This finding aids early-stage drug design by improving compound selection and reducing development attrition.
Area of Science:
- Pharmacokinetics
- Drug Discovery
- Computational Chemistry
Background:
- Pharmacokinetics is crucial for drug design, involving absorption, distribution, metabolism, and excretion (ADME) prediction.
- Accurate prediction of intestinal absorption is vital to reduce drug candidate attrition in preclinical and clinical trials.
- Machine learning approaches are increasingly used to develop predictive models for compound absorption.
Purpose of the Study:
- To evaluate and compare the performance of six artificial intelligence (AI) methods for predicting human intestinal absorption.
- To identify the most effective AI model for accurate prediction of compound intestinal absorption.
Main Methods:
- Six AI methods were employed: Support Vector Machine (SVM), k-Nearest Neighbor (k-NN), Probabilistic Neural Network (PNN), Artificial Neural Network (ANN), Partial Least Squares (PLS), and Linear Discriminant Analysis (LDA).
- These models were trained and validated using datasets to predict the intestinal absorption of compounds.
Main Results:
- SVM achieved the highest prediction accuracy at 91.54%.
- k-NN (88.33%), ANN (86.51%), PNN (84.30%), LDA (80.08%), and PLS (79.07%) followed in accuracy.
- SVM demonstrated superior performance in classifying compounds based on human intestinal absorption.
Conclusions:
- Support Vector Machine, particularly with a Radial Basis Function (RBF) kernel, is the most effective AI method for predicting human intestinal absorption.
- This AI-driven approach can significantly benefit early-stage drug design and development by improving compound selection and reducing failure rates.