Mode of action classification of chemicals using multi-concentration time-dependent cellular response profiles.

In this paper, we present a new statistical pattern recognition method for classifying cytotoxic cellular responses to toxic agents. The advantage of the proposed method is to quickly assess the toxicity level of an unclassified toxic agent on human health by bringing cytotoxic cellular responses with similar patterns (mode of action, MoOA) into the same class. The proposed method is a model-based hierarchical classification approach incorporating principal component analysis (PCA) and functional data analysis (FDA). The cytotoxic cell responses are represented by multi-concentration time-dependent cellular response profiles (TCRPs) which are dynamically recorded by using the xCELLigence real-time cell analysis high-throughput (RTCA HT) system. The classification results obtained using our algorithm show satisfactory discrimination and are validated using biological facts by examining common chemical mechanisms of actions with treatment on human hepatocellular carcinoma cells (HepG2).