Datamining in non-target chemical screening data

In the last 10 years non-target chemical screening data of water samples have been collected at KWR. The liquid chromatography-high resolution mass spectrometry (LC-HRMS) analysis is a more comprehensive method for broad screening water quality than the commonly used target analysis (Ter Laak., 2010) and provides qualitative information on a broad spectrum of chemicals present in the water samples. However, the identification and quantification of all observed compounds remains challenging.
The collected large set of water quality monitoring data contains much information on the presence of chemicals. In the last years, over 150 samples were analysed of waste water effluent, surface water, groundwater and drinking water samples all over The Netherlands. Besides the presence of chemicals, this large dataset contains information on the occurrence of chemicals related to geographical setting, hydrology and land use. However, uncertainties on identification and quantification of compounds make it hard to distil relevant information and obtain knowledge on water quality (Hogenboom et al., 2009).
Therefore, in this study we extract information from the derived data in non-target chemical screening. The study is divided in two phases. The first phase deals with the selection of potential candidates for identification of chemicals using suspect screening. The second phase links the occurrence of compounds to different water types.