|
|
| Type |
Poster Presentation |
| Area |
Medicinal Chemistry |
| Room No. |
Exhibition Hall 2+3 |
| Time |
10월 19일 (목요일) 11:00~12:30 |
| Code |
MEDI.P-329 |
| Subject |
Predicting acute oral toxicity of chemicals by QSAR approach |
| Authors |
JaeYong Lee, Byeong Hun Lee, Sung Kwang Lee*
Department of Chemistry, Hannam University, Korea
|
| Abstract |
Oral toxicity is the toxicity of drugs that enter the digestive tract due to ingestion or absorption, impairing the function of the organism and altering organs or tissue. Approximately 2,000 chemical substances enter the world market each year, and more than 300 kinds of chemicals enter the market every year in Korea. Therefore, a toxicity tests are required to evaluate the toxicity of the compound to prevent adverse effects on human health or the environment. Unfortunately, toxicity test are time and cost consuming, and poses an ethical problem. The use of in silico method based on computer tools to solve these problems is fast and cost-effective to test chemical toxicity as an ethical alternative. The approach to achieving this goal successfully is the quantitative structure-activity relationship(QSAR), which use the mathematical expression of chemical structure information.
In this study, LD50 data according to OECD Guideline 401 were collected from eChemPortal website to predict QSAR model. The molecular descriptors was calculated from 2D chemical structures using the PreADMET program. The data set is divided into training set(60%) used in the model development process and external test set(40%) used in the model validation process. 2D chemical descriptors and machine learning methods such as multiple linear regression(MLR) and support vector machine(SVM) were used to develop predictive QSAR model. The chance correlation and model predictability was verified by y-scrambling method and external validation. Also reliable prediction range is set in the kNN-based applicability domain. |
| E-mail |
inpisky@naver.com |
|
120th General Meeting of the KCS