(274d) Data Mining of High Throughput Screening Toxicity of Engineered Nanoparticles
Nanomaterials are manufactured for their unique properties that enable new applications and products. These specific properties may lead to different interactions with and impacts on ecological receptors and the environment, resulting in effects that can be significantly different from those known for bulk materials. In this regard, in vitro and in vivo toxicity screenings provide valuable information to characterize the effects of nanoparticles (NPs) on biological receptors. The use of high throughput screening (HTS) techniques, however, results in high dimensional datasets whose interpretation requires the use of advanced statistical data analysis and pattern recognition techniques. In this work we adapt and demonstrate the Self-Organizing Map (SOM) algorithm as a tool for knowledge extraction from HTS data sets.
The data used in this study are based on the oxidative stress paradigm and consisted of HTS in vitro analysis of the cytotoxic effects produced by different metal and metal oxide nanoparticles (Pt, Al2O3, CeO2, Co3O4, CuO, SiO2, TiO2, ZnO) on murine RAW 264.7 macrophage cells and human epithelial BEAS-2B cells, respectively. Toxic activity at different exposure times in the range 1 to 6 h was measured using epifluorescence microscopy. The concentration of NPs was in the range of 1 - 200 μg/mL along the screening plate. Fluorescence data were standardized using Z-scores and normalized with respect to untreated samples (i.e., cells not exposed to NPs).
The resulting cytotoxicity profiles were processed using a SOM configured as a hexagonal lattice. The analysis of the clustering structure of the data clearly revealed the presence of well-defined clusters indicating similarity of toxicity patterns. Compared with other commonly used methods such as Heat-Maps and hierarchical clustering, the use of SOM projections facilitates the data mining process since it preserves the preexisting topological relations. The implications of the present approach to developing quantitative-structure-activity relations (QSARs) to predict nanotoxicity of nanoparticles will be presented with the goal of establishing a systematic approach to developing both quantitative and classifier-based QSARs