改进的整合加和模型INFCIM及其应用于混合物毒性预测

目前,准确预测混合物毒性仍然面临着挑战, 为改进现有整合加和模型INFCIM,将该模型中“浓度=浓度+效应”形式修改为更加科学合理的“浓度=浓度+浓度”形式.利用分子电性距离矢量(MEDV)表征混合物组分的分子结构,以模糊数学中的隶属函数表征混合物组分的相似性和相异性,从而构建新的整合加和模型.利用6组六元混合物(共72个样本)验证模型的预测能力,结果表明,改进的模型能够准确预测无相互作用混合物毒性.在改进的模型中,利用多组混合物作为校正集,克服了INFCIM模型仅使用少量混合物数据作为校正集的缺点,使之更加可靠和具有代表性.

Improved integrated addition model INFCIM and its application on prediction of mixture toxicity

Recently, the accurately prediction of mixture toxicity remains a challenge. In order to modify the existed integrated addition model INFCIM (integrated fuzzy concentration addition-independent action model), the form of “concentration = concentration + effect” for the INFCIM model was modified into a reasonable form of “concentration = concentration + concentration”. The molecular electronegativity distance vector was used to characterize the molecular structures of mixture components. The fuzzy set theory was used to describe the degree of similarity and dissimilarity of mixture components. A new integrated addition model was then developed. Six mixtures (including 72 samples) with six components were used to test the predictive ability of the modified model. The results show that the modified model can accurately predict the non-interactive mixture toxicity. The proposed model based on the multiple mixtures overcomes the disadvantage of the model that only uses a single mixture data as calibration set. Thus, the modified model is more reliability and representativeness than the model based on a single mixture.