Recognition of chemical compounds in contaminated water using time-dependent multiple dose cellular responses

An early determination of toxicant compounds of water contaminations can gain critical time to protect citizens' health and save substantial amounts of medical costs. To determine toxins in real time, a multi-dose classification algorithm using cellular state variable identification (CSVID) is developed in this paper. First, the dynamic cytotoxicity response profiles of living cells are measured using a real-time cell electronic sensing (RT-CES) system. Changes in cell number expressed as cell index (CI) are recorded on-line as time series. Then CSVID, which reflects the cell killing, cell lysis and certain cellular pathological changes, is extracted from those dynamic cellular responses. Finally, a support vector machine (SVM) algorithm based on CSVID is employed to classify chemical compounds and determine their analogous cellular response pathway. In order to increase the classification accuracy, a majority vote of the class labels is also proposed. Several validation studies demonstrate that CSVID-based classification algorithm has great potential in distinguishing the cytotoxicity response of the cells in the presence of toxins.