基于QSAR预测模型的有机磷类农药急性毒性的健康风险评估

Risk assessment of acute exposure of the population to organophosphorus pesticides based on QSAR model of toxicity prediction

目的本研究旨在构建一个基于Matlab平台的BP神经网络模块的毒性预测模型,为突发公共卫生事件中缺乏毒理学信息的化学物质的风险评估与预测实践提供科学依据和参考。方法选择20种常见有机磷农药,其中17个为训练集,3个为验证集,预测有机磷农药对硫磷对乙酰胆碱酯酶活性50%抑制浓度。结果考虑不确定因素后,获得对硫磷的急性参考剂量(ARfD)为1.27μg/kg。结论根据2000~2001年中国食品对硫磷含量,采用本模型评估认为通过这些食品摄入对硫磷量的健康风险是可接受的,但对于多种有机磷农药残留的食品,其健康危害风险值得关注。

Objective To develop a toxicity prediction model,based on BP neural network module of Matlab platform,so as to provide scientific basis and reference for the risk assessment of chemicals lacking toxicological information in case of public health emergency.Methods Twenty common organophosphorus pesticides were selected,among which 17 were in training sets and 3 were in validation sets.The 50%inhibitory concentration of parathion on acetylcholinesterase activity was predicted.Results The acute reference dose(ARfD)of parathion was 1.27μg/kg after considering the uncertainties.Conclusion According to the content of parathion in Chinese food monitored from 2000 to 2001,the intake of parathion is considered to be acceptable,however,in view of food with variety of organophosphorus residues,the health risks of which should be paid close attention to.