Prediction of Human Acute Toxicity and Aquatic Toxicity from Chemical Structure by Pattern Recognition
| Project/Area Number |
61571034
|
|---|
| Research Category |
Grant-in-Aid for General Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
Physical pharmacy
|
|---|
| Research Institution | Kitasato University |
|---|
Principal Investigator |
MORIGUCHI Ikuo School of Pharmaceutical Sciences, Kitasato University, 薬学部, 教授 (90050343)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
MATSUSHITA Yasuo School of Pharmaceutical Sciences, Kitasato University, 薬学部, 講師 (40050653)
HIRONO Shuichi School of Pharmaceutical Sciences, Kitasato University, 薬学部, 講師 (30146328)
|
|---|
| Project Period (FY) |
1986 – 1988
|
| Project Status |
Completed (Fiscal Year 1988)
|
| Budget Amount *help |
¥1,700,000 (Direct Cost: ¥1,700,000)
Fiscal Year 1988: ¥200,000 (Direct Cost: ¥200,000)
Fiscal Year 1987: ¥300,000 (Direct Cost: ¥300,000)
Fiscal Year 1986: ¥1,200,000 (Direct Cost: ¥1,200,000)
|
|---|
| Keywords | toxicity prediction / human acute toxicity / aquatic toxicity / structure-toxicity relationship / quantitative structure-activity relationship / pattern recognition / 適応最小二乗法 / 急性毒性 / ヒト経口急性毒性 |
|---|
| Research Abstract |
Quantitative structure-tixicity models have been developed for the prediction of human acute eixicity and gquatic tixicity ratings for a wide variety of organic chemicals using the adaptive least-squares and fuzzy adaptive least-squares.
The discriminant functions are based on substructural fragments obtained for 504 compounds in human acute tixicity and 329 compounds in aquatic toxicity. The data used for the derivation of the functions were mainly obtained from the clinical toxicology of commercial products compiled by gosselin et al. And the registry of Toxic Effects of Chemical Substances of U.S.Department of Health and Human Services.
The acxcuracy of classifying the chemicals into three toxicity ratings was 87% in the recognition and 82% in the leave-one out prediction for human acute toxicity, and 85% in the recognition and 78% in the leave-one-out prediction for aquatic toxicity. It is evident from these results that reasonably accurate discrimanant functions can be designed for the estimation of human and aquatic toxicity ratings.
|
Report
(4 results)
Research Products
(14 results)
