Elsevier

Chemical Physics Letters

Volume 660, 1 September 2016, Pages 107-110
Chemical Physics Letters

Research paper
Quasi-SMILES and nano-QFPR: The predictive model for zeta potentials of metal oxide nanoparticles

https://doi.org/10.1016/j.cplett.2016.08.018Get rights and content

Highlights

  • The predictive model for zeta potentials is developed.

  • The predictive value is calculated with quasi-SMILES.

  • In contrast to traditional SMILES quasi-SMILES is representation of conditions.

  • Each condition is represented by a code.

  • Optimal descriptor is sum of correlation weights of the codes of conditions.

  • Numerical data on the correlation weights are calculated with the Monte Carlo method.

Abstract

Building up of the predictive quantitative structure–property/activity relationships (QSPRs/QSARs) for nanomaterials usually are impossible owing to the complexity of the molecular architecture of the nanomaterials. Simplified molecular input-line entry system (SMILES) is a tool to represent the molecular architecture of “traditional” molecules for "traditional" QSPR/QSAR. The quasi-SMILES is a tool to represent features (conditions and circumstances), which accompany the behavior of nanomaterials. Having, the training set and validation set, so-called quantitative feature–property relationships (QFPRs), based on the quasi-SMILES, one can build up model for zeta potentials of metal oxide nanoparticles for situations characterized by different features.

Introduction

Quantitative structure–property/activity relationships (QSPR/QSAR) based on different descriptors are a tool to build up predictive model for endpoints of different substances as a mathematical function of their molecular structure represented by the molecular graph [1]. Simplified molecular input-line entry system (SMILES) is a possible alternative of molecular graph for representation of the molecular structure for the QSPR/QSAR [2], [3]. The CORAL software [4] gives possibility to build up QSPR/QSAR models where the molecular structure is represented by SMILES [5], [6], [7].

Intensive research work on the nanomaterials stimulates the search for approaches aimed to predict physicochemical and biochemical behavior of nanomaterials [8].

However, the ‘traditional’ QSPR/QSAR analysis can solve not all tasks related to nanomaterials, because

  • (i)

    The limited number of ‘more or less’ regular sources of data on nanomaterials are available for praxis;

  • (ii)

    Very complex molecular structure of nanomaterials, as a rule, cannot be represented by graph or SMILES; and

  • (iii)

    Usually, a physicochemical and biochemical experiments are based on analysis of conditions (dose, irradiation, time of exposure, etc.), in other words, the molecular structure of nanomaterials sometimes has no influence on an experimental result.

The quasi-SMILES [9], [10], [11], [12], [13], [14] is possible way to build up predictive models for nanomaterials. In contrast to traditional SMILES, quasi-SMILES are representations of conditions. It is to be noted the molecular structure in principle can be examined as a special kind of conditions, if it is expedient [15].

The development of quantitative feature-property relationships (QFPRs), based on quasi-SMILES, for zeta potentials of metal oxide nanoparticles is aim of this study.

Section snippets

Data

The numerical data on the zeta potential [mV] of metal oxide nanoparticles are taken from the literature [16]. The model is a mathematical function of different features of nanoparticles. The features are first, fifteen metal oxides, and second, four circumstances: (i) Acid (pH 5.6); (ii) Basic (pH 7.4); (iii) Serum corona; and (iv) Surfactant corona. Table 1 contains the list of the features above and their representations in quasi-SMILES. The quasi-SMILES were randomly split into the training

Results and discussion

Table 2 contains the models of zeta potentials obtained with three different splits into the training, calibration, and validation sets. One can see, the statistical characteristics of models for different splits are comparable and enough good (Table 2).

Table 3 contains the correlation weights for three random splits into the training, calibration, and validation sets, which are utilized to calculate the DCW (T∗, N∗).

Table 4 contains the numerical data on the experimental and calculated zeta

Conclusions

The quasi-SMILES (Table 1) gives possibility to build up predictive model for zeta potential of metal oxide nanoparticles in the form of quantitative feature-property relationships (QFPRs) similar to described early QFAR. The approach gives models for zeta potential of metal oxide nanoparticles according to OECD principles.

Acknowledgments

AAT and APT thank the EC project PeptiCAPS (Project reference: 686141).

Cited by (31)

  • Nano-QSAR modeling for predicting the cytotoxicity of metallic and metal oxide nanoparticles: A review

    2022, Ecotoxicology and Environmental Safety
    Citation Excerpt :

    Therefore, it is important for the construction of nano-QSAR models to identify a new set of descriptors that can accurately represent the characteristics of NPs as well as the experimental conditions. In this context, the simplified molecular input-line entry (SMILE) system or quasi-SMILE method (Fig. 2a), which refers to the conversion of chemical structures into a sequence of symbols for computer recognition and analysis, has been proposed by several researchers for modeling because it can represent molecular structures, physiochemical properties, and exposure conditions (Toropov et al., 2016; Toropova and Toropov, 2017). Toropova et al. (2014) used the quasi-SMILE method to predict the membrane damage caused by a group of ZnO and TiO2 NPs.

View all citing articles on Scopus
View full text