Stenberg Mia, Linusson Anna, Tysklind Mats, Andersson Patrik L
Department of Chemistry, Umeå University, SE-901 87 Umeå, Sweden.
Chemosphere. 2009 Aug;76(7):878-84. doi: 10.1016/j.chemosphere.2009.05.011. Epub 2009 Jun 9.
In present study the Industrial chemical map was created, and investigated. Molecular descriptors were calculated for 56072 organic substances from the European inventory of existing commercial chemical substances (EINECS). The resulting multivariate dataset was subjected to principal component analysis (PCA), giving five principal components, mainly reflecting size, hydrophobicity, flexibility, halogenation and electronical properties. It is these five PCs that form the basis of the map of organic, industrial chemicals, the Industrial chemical map. The similarities and diversity in chemical characteristics of the substances in relation to their persistence (P), bioaccumulation (B) and long-range transport potential were then examined, by superimposing five sets of entries obtained from other relevant databases onto the Industrial chemical map. These sets displayed very similar diversity patterns in the map, although with a spread in all five PC vectors. Substances listed by the United Nations Environment Program as persistent organic pollutants (UNEP POPs) were on the other hand clearly grouped with respect to each of the five PCs. Illustrating similarities and differences in chemical properties are one of the strengths of the multivariate data analysis method, and to be able to make predictions of, and investigate new chemicals. Further, the results demonstrate that non-testing methods as read-across, based on molecular similarities, can reduce the requirements to test industrial chemicals, provided that they are applied carefully, in combination with sound chemical knowledge.
在本研究中，创建并研究了工业化学品图谱。从欧洲现有商业化学物质清单（EINECS）中计算了56072种有机物质的分子描述符。对所得的多变量数据集进行主成分分析（PCA），得到五个主成分，主要反映大小、疏水性、柔韧性、卤化和电子性质。正是这五个主成分构成了有机工业化学品图谱——工业化学品图谱的基础。然后，通过将从其他相关数据库获得的五组条目叠加到工业化学品图谱上，研究了这些物质在持久性（P）、生物累积性（B）和远距离迁移潜力方面化学特征的相似性和多样性。尽管在所有五个主成分向量中都有分布，但这些组在图谱中显示出非常相似的多样性模式。另一方面，联合国环境规划署列为持久性有机污染物（UNEP POPs）的物质在五个主成分中的每一个方面都明显聚类。说明化学性质的异同是多变量数据分析方法的优势之一，并且能够对新化学品进行预测和研究。此外，结果表明，基于分子相似性的“类推法”等非测试方法，如果谨慎应用并结合扎实的化学知识，可以减少对工业化学品进行测试的要求。
