A machine learning appraoch to data-driven discovery of drug molecules
| Project/Area Number |
25540015
|
|---|
| Research Category |
Grant-in-Aid for Challenging Exploratory Research
|
| Allocation Type | Multi-year Fund |
|---|
| Research Field |
Statistical science
|
|---|
| Research Institution | The Institute of Statistical Mathematics |
|---|
Principal Investigator |
RYO Yoshida 統計数理研究所, モデリング研究系, 准教授 (70401263)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
IBA Yukito 統計数理研究所, モデリング研究系, 教授 (30213200)
|
|---|
| Project Period (FY) |
2013-04-01 – 2015-03-31
|
| Project Status |
Completed (Fiscal Year 2014)
|
| Budget Amount *help |
¥3,640,000 (Direct Cost: ¥2,800,000、Indirect Cost: ¥840,000)
Fiscal Year 2014: ¥1,820,000 (Direct Cost: ¥1,400,000、Indirect Cost: ¥420,000)
Fiscal Year 2013: ¥1,820,000 (Direct Cost: ¥1,400,000、Indirect Cost: ¥420,000)
|
|---|
| Keywords | 化学情報学 / 創薬 / 分子設計 / ベイズ統計 / カーネル法 / マルコフ連鎖モンテカルロ法 |
|---|
| Outline of Final Research Achievements |
The chemical space subject to pharmaceutical developments consists of 10~60 pharmaceutical candidates. Our challenge is to discover new compounds in the huge space that exhibit desired properties on various pharmacological activities, required to be a drug. The aim of this study is to create a new molecular design method by the integration of Bayesian and machine learning methods. Our approach is as follows: (a) we develop forward prediction models based on experimental data that predict the properties of a compound with the chemical structure, (ii) the backward prediction is derived by inverting the forward model according to the Bayes formula, and (iii) the backward model is used to generate new compounds with the chemical structures likely to achieve desired pharmacological activities. Under industry-academia partnerships, we are putting into practice the developed method in the area of resin and pigment chemistry in addition to pharmaceutical developments.
|
Report
(3 results)
Research Products
(5 results)
