Edge Effect Detection for Real-time Cellular Analyzer using Functional Principal Component Analysis

To detect cytotoxicity of chemicals, many instruments have been developed. One popular tool is real time cellular analyzer (RTCA). Nevertheless, abnormal time-dependent cellular response curves (TCRCs) always occur and disturb experimental results when the wells are at the edge of E-plate. Therefore, a method is proposed to detect edge effect which is detrimental to the experimental quality. In this work, these TCRCs were considered as observations of a random variable on a functional space and we utilized Functional Principal Component Analysis (FPCA) to extract principal components of TCRCs to find unusual curves. The average normalized cell index (NCI) of the inner wells was defined as the standard. Then all TCRCs were analyzed by FPCA to find abnormal TCRCs which would be removed automatically by computer. This approach has never been applied in RTCA system to determine edge effect. Experimental results indicate that the FPCA algorithm achieves a comparable detection rate.