| Presentation | 2008-06-26 Prediction of Drug Clearance Pathway with Machine Learning Kouta TOSHIMOTO, Makiko KUSAMA, Kazuya MAEDA, Yuichi SUGIYAMA, Yutaka AKIYAMA, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | The clearance pathway is one of the important factors to consider the pharmacokinetics of drugs. We have developed a machine learning system of drug clearance pathway for a given drug compound from its physicochemical descriptors. The system is composed of five support vector machines (SVMs), each corresponding to one of five major candidate clearance pathways, and prediction is given by choosing a pathway with largest SVM output. We prepared pathway data for 157 drugs, and 1089 physicochemical descriptors for each of them. However, if we use all the descriptors, we will have over-learning problem and less explainable model. Thus we performed exhaustive feature selection procedure, by a modified greedy algorithm or a correlation coefficient-based method, and our system showed more than 85% prediction accuracy when using 12 selected descriptors. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Machine learning / Pharmacokinetics / Support vector machine(SVM) / Feature selection |
| Paper # | NLP2008-4,NC2008-14 |
| Date of Issue |
| Conference Information | |
| Committee | NC |
|---|---|
| Conference Date | 2008/6/19(1days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | |
| Vice Chair | |
| Secretary | |
| Assistant | |
| Paper Information | |
| Registration To | Neurocomputing (NC) |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Prediction of Drug Clearance Pathway with Machine Learning |
| Sub Title (in English) | |
| Keyword(1) | Machine learning |
| Keyword(2) | Pharmacokinetics |
| Keyword(3) | Support vector machine(SVM) |
| Keyword(4) | Feature selection |
| 1st Author's Name | Kouta TOSHIMOTO |
| 1st Author's Affiliation | Graduate School of Information Science and Engineering, Tokyo Institute of Technology() |
| 2nd Author's Name | Makiko KUSAMA |
| 2nd Author's Affiliation | Graduate School of Pharmaceutical Sciences, The University of Tokyo |
| 3rd Author's Name | Kazuya MAEDA |
| 3rd Author's Affiliation | Graduate School of Pharmaceutical Sciences, The University of Tokyo |
| 4th Author's Name | Yuichi SUGIYAMA |
| 4th Author's Affiliation | Graduate School of Pharmaceutical Sciences, The University of Tokyo |
| 5th Author's Name | Yutaka AKIYAMA |
| 5th Author's Affiliation | Graduate School of Information Science and Engineering, Tokyo Institute of Technology |
| Date | 2008-06-26 |
| Paper # | NLP2008-4,NC2008-14 |
| Volume (vol) | vol.108 |
| Number (no) | 101 |
| Page | pp.pp.- |
| #Pages | 6 |
| Date of Issue | |