|
(461) Multi-criteria prioritisation of xenobiotics. Lerche, Dorte1,2, Sørensen, Peter2, Thomsen, Marianne2,3, 1 2 3 ABSTRACT- Prioritisation of xenobiotics for eco-toxicological impact is a needed but complicated task for decision making. Valid scientific analytical conclusions are hard to obtain. However, multi-criteria prioritisation approaches using partial order ranking theory appear to be a promising tool, being transparent and scientifically well founded. The methodology offers the possibility to address various groups of xenobiotics and to include chemical and biological parameters simultaneously. The interpretations of mullet-criteria prioritisation results are nevertheless hindered when a large number of incomparable sets of xenobiotics is involved. In general the number of incomparable sets increases with the number of independent parameter used. In this study we have investigated the use of linear extensions within partial orders systems to estimate the rank of the incomparable sets of xenobiotics. If all linear extensions are found the most possible ranking can be calculated. In small systems with 10-15 xenobiotics the total number of linear extensions can be found easily, but in larger systems the number of possible combinations makes this impossible. This motivated the development and investigation of two Monte Carlo simulations. In the first approach a computer program generates random linear extensions in a partial order system. Based on the randomly chosen linear extensions the most possible rank can be estimated. The second approach iterated the most possible rank for each set of incomparable xenobiotics. The results from the two methods are compared. Further, QSAR data on log Kow and EC50 for four groups of xenobiotics (phthalate esters, PCBs, PAHs and pesticides) were used to illustrate the methodology. Key words: prioritization, multi-criteria, partial order, QSAR |
