Mining a Nanoparticle Dataset, Compiled Within the MODENA-COST Action

Engineered nanomaterials (ENM) have new or enhanced physico-chemical properties compared to their micron-sized counterparts, but may also have an increased toxic potential. Animal and in vitro testing are typically employed to investigate the toxic effects of (nano)materials. The sheer number of ENMs and their physico-chemical parameters make it impossible to only use in vivo and in vitro testing, and modelling technologies are also deployed to find relationships between ENM parameters and toxicity. A heterogenous dataset containing information on 192 nanoparticle endpoints was compiled within the MODENA COST-Action consortium. Here, the available data was mined to identify relationships between nanoparticle properties and cell-death as measured with four cytotoxicity assays. ANOVA, collinearity analyses and classification and regression trees gave indications on potential relations between the NP-properties and toxicity, but could not deliver a robust model. More information and datapoints are necessary to build well-validated models.