Identifying diverse metal oxide nanomaterials with lethal effects on embryonic zebrafish using machine learning

• Richard Liam Marchese Robinson,
• Haralambos Sarimveis,
• Philip Doganis,
• Xiaodong Jia,
• Marianna Kotzabasaki,
• Christiana Gousiadou,
• Stacey Lynn Harper and
• Terry Wilkins

Beilstein J. Nanotechnol. 2021, 12, 1297–1325, doi:10.3762/bjnano.12.97

• be uncoated), it was necessary to assign dummy values for the molecular descriptors where the corresponding components did not exist (i.e., where the descriptors were not applicable). However, the dataset was selected such that, in contrast to non-applicable dummy values [45], there were no missing

• “survival = 100%”, respectively. Molecular descriptors were computed as per the treatment of ENM surface components and dummy values were assigned for all other descriptors. The similarity weighting of these samples was treated as a tunable hyperparameter between 0.1 and 0.9. (2) The new samples were

• , was represented using binary variables for each value. Their negligible significance might partly reflect the limited availability of this information. Finally, the specific chemical composition of the organic surface coatings was represented using molecular descriptors designed to capture

Evaluating the toxicity of TiO₂-based nanoparticles to Chinese hamster ovary cells and Escherichia coli: a complementary experimental and computational approach

• Alicja Mikolajczyk,
• Natalia Sizochenko,
• Ewa Mulkiewicz,
• Anna Malankowska,
• Michal Nischk,
• Przemyslaw Jurczak,
• Seishiro Hirano,
• Grzegorz Nowaczyk,
• Adriana Zaleska-Medynska,
• Jerzy Leszczynski,
• Agnieszka Gajewicz and
• Tomasz Puzyn

Beilstein J. Nanotechnol. 2017, 8, 2171–2180, doi:10.3762/bjnano.8.216

• approach is based on defining mathematical dependencies between the variance in molecular structures, encoded by so-called molecular descriptors, and the variance in a given physicochemical property or biological (e.g., cytotoxicity) property in a set of compounds (“endpoints”) [5][25][26][27][28][29][30

Modeling adsorption of brominated, chlorinated and mixed bromo/chloro-dibenzo-p-dioxins on C₆₀ fullerene using Nano-QSPR

• Piotr Urbaszek,
• Agnieszka Gajewicz,
• Celina Sikorska,
• Maciej Haranczyk and
• Tomasz Puzyn

Beilstein J. Nanotechnol. 2017, 8, 752–761, doi:10.3762/bjnano.8.78

• was generated as a part of the Persistent Organic Pollutants Big Data project by using ConGENER software [47], and described in more detail in our previous study [48]. A set of 26 so-called molecular descriptors was calculated for each congener at the semi-empirical PM6 level. A descriptor, by

• Information File 1. Holland’s genetic algorithm (GA) [61] was used for the selection of the optimal combination of molecular descriptors and redundancy elimination in the structural data. Partial least squares (PLS) regression was applied as the method of modeling to solve the common problem of co-linearity

Nanoinformatics for environmental health and biomedicine

• Rong Liu and
• Yoram Cohen

Beilstein J. Nanotechnol. 2015, 6, 2449–2451, doi:10.3762/bjnano.6.253

• components calculated from nanoparticle size and surface properties using Kriging estimations [14]. Another contribution reports on the development of models to predict the cytotoxicity of PAMAM dendrimers using molecular descriptors [15]. Nanomaterials that have potential to cause disease (e.g., TiO2

Predicting cytotoxicity of PAMAM dendrimers using molecular descriptors

• David E. Jones,
• Hamidreza Ghandehari and
• Julio C. Facelli

Beilstein J. Nanotechnol. 2015, 6, 1886–1896, doi:10.3762/bjnano.6.192

• journal articles. The results indicate that data mining and machine learning can be effectively used to predict the cytotoxicity of PAMAM dendrimers on Caco-2 cells. Keywords: data mining; machine learning; molecular descriptors; poly(amido amine) dendrimers (PAMAM); Introduction In silico approaches

• potentially be expanded to other nanomaterials in the future. Results and Discussion Five different analyses were performed to classify a dendrimer as toxic or nontoxic using different combinations of molecular descriptors and experimental conditions. The first analysis utilized all the molecular descriptors

• available in MarvinSketch (see Experimental section and Table S1 in Supporting Information File 1). The second analysis involved an automatic feature selection method in which the molecular descriptors that were used had a nonzero rank according to the ChiSquaredAttributeEval method in Weka (see details in