Additional file 2 of A Nano-QSTR model to predict nano-cytotoxicity: an approach using human lung cells data

Additional file 2. Figure S1 ENMs diameter (nm) frequency distribution (a) before and (b) after the filtering step. Figure S2 Cell viability (%) frequency distribution (a) before and (b) after the filtering step. Figure S3 General representation of data normalization and transformation pre-processing steps for ENMs diameter and concentration. Figure S4 Dataset overview. (a), (b), and (c) Diversity of a set of nano-descriptors obtained from the Elemental descriptor calculator; (e), (f), and (g) Cell Viability (%) variation trend with representative examples of each nano-descriptor. Figure S5 Learning curves for ET nano-QSTR model for training and validation sets according to (a) determination coefficient (R2) and (b) root-mean-square error (RMSE). Table S1 ET nano-QSTR model performance for training, validation, test, and unseen sets assuming data normalization by ENMs surface area.