Today, predictive methods like Quantitative Structure-Activity Relationships (QSARs) play an important role in setting the priorities for risk assessment of chemicals. However, when a large number of compounds (>200) have to be considered, limitations arise on which physico-chemical variables to use. In this paper, a fast and easy way of chemically characterising compounds has been investigated. A QSAR is a regression model that relates the variation in the biological activity (y-matrix) within a homogenous series of compounds to their chemical and structural properties (x-matrix). The descriptor variable matrix (x) may be constructed from a multitude of variables indicating either the presence or absence of predefined structural fragments. The present paper investigates the possibilities and limitations of using such a derived x-matrix in cases of QSARs and environmental modelling. The approach is used in two examples, which both gave positive results.
Wahlstöm B.: The Need for New Strategies- the OECD Existing Chemicals Program; in: The Use of QSAR for Chemicals Screening- Limitations and Possibilities, Report No. 8. National Chemicals Inspectorate, Stockholm, 1988.
2.
Blum Djw, Speece RE: Determining Chemical Toxicity to Aquatic Species. Environ Sci Technol1990;24:284-293.
3.
Directorate-General XI for Environment, Nuclear Safety and Civil Protection, Chemical Risk Control. Commission of the European Communities, Luxembourg , 1994.
4.
Priority setting using the IPS method: From IUCLID to IPS ranking. Commission of the European Communities, Joint Research Centre, European Chemicals Bureau, Ispra, doc. XI/ECB/883/93 (rev. 1).
5.
Eriksson L.: A Strategy for Ranking Environmentally Occurring Chemicals; PhD thesis, Umeå University, Umeå1991.
6.
Nendza M., Volmer J., Klein W.: Risk assessment based on QSAR estimates; in Devillers J, Karcher W (eds): Practical Applications of Quantitative Structure-Activity Relationships (QSAR) in Environmental Chemistry and Toxicology. Dordrecht, Kluwer, 1990, pp 213-240.
7.
American Chemical Society: CAS on-line Screen Dictionary for Substructure Search, ed 2. Columbus, Ohio, 1981.
8.
James CA, Weininger D.: Daylight Theory Manual. Daylight Chemical Information Systems , Irvine, 1994.
9.
Lajiness MS: Molecular similarity-based methods for selecting compounds for screening; in Rouvray DH (ed): Computational Chemical Graph Theory. New York, Nova Science Publishers, 1990, pp 299-316.
10.
Martin EJ, Blaney JM, Siani MA, Spellmeyer DC, Wong AK, Moos WH: Measuring diversity: Experimental design of combinatorial libraries for drug discovery. J Med Chem1994; 38:1431-1436.
11.
Downs GM, Wilett P.: The use of similarity and clustering, techniques for the prediction of molecular properties; in Devillers J, Karcher W (eds): Applied Multivariate Analysis in SAR and Environmental Studies. Dordrecht, Kluwer, 1991, pp 247-279.
12.
Barnard JM, Downs GM: Clustering of chemical structures on the basis of two-dimensional similarity measures. J Chem Inf ComputSci1992;32:644-649.
13.
Downs GM, Gillet VJ, Holliday JD, Lynch MF: Computer storage and retrieval of generic chemical structures in patients. 10. Assignment and logical bubble-up of ring screens for structurally explicit generics. J Chem Inf ComputSci1989;29:215-224.
14.
Eakin DR; The IciCROSSBOW System; in Ash JE, Hyde E (eds): Chemical Information Systems . Chichester, Ellis Horwood, 1974, pp 227-241.
15.
Wold S., Martens H., Wold H.: The Multivariate Calibration Problem in Chemistry Solved by the PLS Method ; in Ruhe A, Kågström B (eds): Matrix Pencils. Heidelberg , Springer, 1983, pp 286-293.
16.
Wold S., Ruhe A., Wold H., Dunn WJIII: The collinearity problem in linear regression. The partial least squares approach to generalized inverses. SIAM J Sci Stat Comput1984 ;5:735-743.
17.
Geladi P., Kowalski BR: Partial least squares regression (PLS): A tutorial. Anal Chim Acta1986;185:1-17.
18.
Manne R.: Analysis of two partial least squares algorithms for multivariate calibration . Chemometr Intell Lab Syst1987;2:187-197.
19.
Helland IS: The structure of partial least squares regression. Commun Stat Simul Comput1988;17:581-607.
Garthwaite PH: An interpretation of partial least squares . J Am Stat Assoc1994 ;89:122-127.
22.
Lindgren F.: Third Generation PLS: Some elements and applications; PhD thesis, Umeå University, 1994.
23.
Stone M.: Cross-validatory choice and assessment of statistical predictions. J R Stat Soc1974;36:111-133.
24.
Geisser S.: A predictive approach to the random effect model. Biometrika1974;61:101-107.
25.
Wold S.: Cross-validatory estimation of the number of compounds in factor and principal components models. Technometrics1978;20: 379-406.
26.
Bol J., Verhaar Hjm, Van Leeuwen CJ, Hermens Jlm: Report prepared within the framework of the project 'Predictions of the Aquatic Toxicity of High-Production-Volume Chemicals Part B Predictions.' A project of the Ministry of Housing, Physical Planning and Environment, Directorate-General for Environmental Protection, contract number SVS 1993/9B.
27.
Verhaar Hjm , Van Leeuwen CJ, Hermens Jlm: Chemosphere1992;25:471-491.
28.
Iuclid; Commission of the European Communities, European Chemicals Bureau, T.P. 280, JRC Ispra Establishment, Ispra , Italy.
