Quantitative Structure-property Relationship Analysis Research Articles

A Hierarchical Hybrid Method for Screening Ionic Liquid Solvents for Extractions Exemplified by the Extractive Desulfurization Process

A hierarchical hybrid method combining experimental-database-derived estimation of extraction performance, quantitative structure–property relationship (QSPR)-based assessment of IL physical and environmental properties, liquid–liquid extraction (LLE) measurement, and process evaluation is proposed to screen practically suitable ionic liquid (IL) solvents for different extractions. From the literature, 47 424 infinite dilution activity coefficient (IDAC) data including 12 IL families (e.g., imidazolium, pyridinium, ammonium, etc.) and 13 organic families (e.g., alkanes, aromatics, alcohols, etc.) are collected. On the basis of the IDAC data, the extraction performance of ILs for a specific separation can be estimated in terms of the distribution ratio and selectivity at infinite dilution. The ILs with potentially high extraction performance and meeting the physical and environmental properties criteria are selected to perform LLE experiments. Subsequently, process simulation and evaluation using the selected IL solvents are performed by Aspen Plus. To exemplify the proposed method, the extractive desulfurization (EDS) process is taken as a case study, where [EMIM][MESO₃] (1-ethyl-3-methylimidazolium methanesulfonate) and [EIM][NO₃] (1-ethylimidazolium nitrate) are selected after IDAC database searching and QSPR analysis. Experimental LLE with the two ILs are determined, demonstrating their promising extraction performance with the maximum selectivity (S₂₃ᵐᵃˣ) for thiophene/heptane of 420 and 281.9, respectively. By fitting the NRTL model correspondingly, two processes using the screened ILs and sulfolane are developed and compared using Aspen Plus. It turns out that the two ILs save 66% and 48% in solvent requirements and 54% and 55% in energy consumption compared to those of sulfolane for the EDS task, respectively.

Industrial Case Study: Identification of Important Substructures and Exploration of Monomers for the Rapid Design of Novel Network Polymers with Distributed Representation

Abstract Designing polymers experimentally is a time-consuming task. Quantitative structure-property relationship analysis can help speed the development of new polymers. The authors hypothesized the ideal mixture model, with which polymers are represented by composition-weighted descriptors of monomers. In this study, we pursued a new polymer that had the desired properties from an industrial dataset. We first constructed a partial least squares (PLS) model and random forest with five descriptor sets. The PLS model with fragment counts, which was the most appropriate model for prediction, was used to optimize the compositions. Subsequently, the authors identified the important substructures of monomers using least absolute shrinkage and selection operator (LASSO). The important substructures were used to select seed structures of monomers for structure generation. Another PLS model with distributed representation, called mol2vec, was constructed, because the ordinary fragment counts are unavailable for extrapolation. The PLS model estimated the polymer target property for screening novel structures. The major novelties of this study are to identify important substructures to the polymer target property and to apply mol2vec to design of network polymers. Eventually, we found a novel desired polymer through the composition optimization and demonstrated that virtual screening of monomers with distributed representation worked.

QSPR analysis of some novel neighbourhood degree-based topological descriptors

Topological index is a numerical value associated with a chemical constitution for correlation of chemical structure with various physical properties, chemical reactivity or biological activity. In this work, some new indices based on neighborhood degree sum of nodes are proposed. To make the computation of the novel indices convenient, an algorithm is designed. Quantitative structure property relationship (QSPR) study is a good statistical method for investigating drug activity or binding mode for different receptors. QSPR analysis of the newly introduced indices is studied here which reveals their predicting power. A comparative study of the novel indices with some well-known and mostly used indices in structure-property modelling and isomer discrimination is performed. Some mathematical properties of these indices are also discussed here.

Systematic study of the structure-property relationship of a series of near-infrared absorbing push-pull heptamethine chromophores for electro-optics

This study reports the facile synthesis, characterization and quantitative structure-property relationship analysis of molecular and material properties of tricyanofuran-based (TCF) dipolar heptamethines with different electron donors of indoline ( F1 ), benzo[ e ]- indoline ( F2 ), benz[ cd ]indoline ( F3 ), and Michler’s base derivatives ( M1 and M2 ). The linear and nonlinear optical (NLO) properties of these chromophores have been thoroughly investigated, and the relationship between molecular and bulk response has been analyzed and compared with dipolar tetraene AJLZ53 as one of the best chromophores for electro-optic (EO) devices. In particular, we provide responsible data collection and analysis of optical and EO properties for poled thin films using a widely-recognized and accredited methodology of prism coupling system with the help of rigid oriented gas model. We found that these push-pull heptamethines with synthetic efficacy exhibit high near-infrared absorption, excellent chemical stability and large hyperpolarizabilities ( β ) varied from 1,023×10−30 esu for F1 , 3,047×10−30 esu for M1 , and 3,547×10−30 esu for F3 at 1,304 nm in poled films, respectively. The β values of these molecules are among the highest ones for TCF-based dipolar chromophores, and also agree well with reported analytical results in the solutions. In poled polymers with a modest chromophoric loading density of 1.3× 1020 cm−3, M1 and M2 give a high poled-induced noncentrosymmetric order and relatively large r 33 values around 40 pm V−1 at 1,304 nm, indicative of large μβ values and suitable structural modification for high poling efficiency. Furthermore, a binary EO polymer based on the co-loading of M1 and AJLZ53 achieve a large r 33 value of 143.3 pm V−1 at 1,304 nm. Our studies suggest that concise synthesis and molecular design of push-pull polymethines can be well guided by the tabulation of their linear and NLO properties in bulk materials, and streamline future development of high performance organic EO materials for photonic applications.

Multiplicative degree based topological indices of some chemical structures in drug

In quantitative structure property relationship analysis (QSPR) and quantitative structure property relationship analysis (QSAR) the correlation between different properties/activities and molecular structure of chemical compounds is investigated which is helpful in drug design. Topological index is an useful tool to predict different physical and chemical properties of molecule by collecting information from the molecular graph. In this article, multiplicative degree based topological indices are obtained for some chemical structures widely used in drug design, especially in anticancer drug discovery. To visualize the indices, the results are interpreted graphically.

Singlet oxygen generation by porphyrins and metalloporphyrins revisited: A quantitative structure-property relationship (QSPR) study

Porphyrins and metalloporphyrins are used as photosensitizers in photocatalysis, photodynamic therapy (PDT), disinfection, degradation of persistent pollutants and other applications. Their mechanism of action involves intersystem crossing to triplet excited state followed by formation of singlet oxygen (1O2), which is a highly reactive species and mediates various oxidative processes. The design of advanced sensitizers based on porphyrin compounds have attracted significant attention in recent years. However, it is still difficult to predict the efficiency of singlet oxygen generation for a given structure. Our goal was to develop a quantitative structure-property relationship (QSPR) model for the fast virtual screening and prediction of singlet oxygen quantum yields for pophyrins and metalloporphyrins. We performed QSPR analysis of a dataset containing 32 compounds, including various porphyrins and their analogues (chlorins and bacteriochlorins). Quantum-chemical descriptors were calculated using Density Functional Theory (DFT), namely B3LYP and M062X functionals. Three different machine learning methods were used to develop QSPR models: random forest regression (RFR), support vector regression (SVR), and multiple linear regression (MLR). The optimal QSPR model «structure – singlet oxygen generation quantum yield» obtained using RFR method demonstrated high determination coefficient for the training set (R2 = 0.949) and the highest predicting ability for the test set (pred_R2 = 0.875). This proves that the developed QSPR method is realiable and can be directly applied in the studies of singlet oxygen generation both for free base porphyrins and their metal complexes. We believe that QSPR approach developed in this study can be useful for the search of new poprhyrin photosensitizers with enhanced singlet oxygen generation ability.

Assessing Graph-based Deep Learning Models for Predicting Flash Point.

Flash points of organic molecules play an important role in preventing flammability hazards and large databases of measured values exist, although millions of compounds remain unmeasured. To rapidly extend existing data to new compounds many researchers have used quantitative structure-property relationship (QSPR) analysis to effectively predict flash points. In recent years graph-based deep learning (GBDL) has emerged as a powerful alternative method to traditional QSPR. In this paper, GBDL models were implemented in predicting flash point for the first time. We assessed the performance of two GBDL models, message-passing neural network (MPNN) and graph convolutional neural network (GCNN), by comparing against 12 previous QSPR studies using more traditional methods. Our result shows that MPNN both outperforms GCNN and yields slightly worse but comparable performance with previous QSPR studies. The average and Mean Absolute Error (MAE) scores of MPNN are, respectively, 2.3 % lower and 2.0 K higher than previous comparable studies. To further explore GBDL models, we collected the largest flash point dataset to date, which contains 10575 unique molecules. The optimized MPNN gives a test data of 0.803 and MAE of 17.8 K on the complete dataset. We also extracted 5 datasets from our integrated dataset based on molecular types (acids, organometallics, organogermaniums, organosilicons, and organotins) and explore the quality of the model in these classes.

Application of Quantitative Structure-Property Relationship (QSPR) Models for the Predictions of Critical Micelle Concentration of Gemini Imidazolium Surfactants

The critical micelle concentration (cmc) is important indexes to determine the performance of surfactants quantitatively. In this study, based on the molecular descriptors, which calculated, by Dragon software, were proposed a quantitative structure-property relationship analysis for prediction of cmc of gemini imidazolium surfactants. In the present study, the associative neural networks (ASNN) and multiple linear regression analysis (MLRA) technique were used to interpret the chemical structural functionality (molecular descriptors) against the cmc of gemini imidazolium surfactants. The models were validated rigorously through 5-fold cross-validation. The ASNN method showed to be better than the MLRA method in terms of the internal and the external prediction accuracy with high statistical quality r2 = 0.96, q2 = 0.95, RMSE = 0.08 and MAE = 0.04 respectively. The developed QSPR models are publicly available on the web site at https://ochem.eu/model/11583111 for ASNN model and https://ochem.eu/model/43858671 for MLRA model. These models can be applied to predict the cmc of new gemini imidazolium surfactants.

Quantitative structure-property relationship (QSPR) analysis of calcium aluminosilicate glasses based on molecular dynamics simulations

Quantitative structure-property relationship (QSPR) analysis of a series of calcium aliminosilicate (CaO-Al2O3-SiO2) glasses (103 compositions) have been performed based on structural information from molecular dynamics (MD) simulations. The definition of the structural descriptor has been explored to find the optimal form for a wide range of properties such as density, Young's modulus, hardness, and thermal expansion coefficient. In particular, different cation-oxygen single bond energies and types of short- and medium-range structural information (e.g., coordination number, network connectivity) used to define the structural descriptor Fnet were explored. The results show slightly different definitions might be needed to achieve desirable correlations between the descriptors and various properties. The work shows that QSPR analysis based on MD generated glass structures is a very promising approach to correlate a wide range of physical properties with glass compositions, hence a very valuable tool for new glass discovery and composition design.

QSPR Analysis with Curvilinear Regression Modeling and Topological Indices

Topological indices are the real number of a molecular structure obtained via molecular graph G. Topological indices are used for QSPR, QSAR and structural design in chemistry, nanotechnology, and pharmacology. Moreover, physicochemical properties such as the boiling point, the enthalpy of vaporization, and stability can be estimated by QSAR/QSPR models. In this study, the QSPR (Quantitative Structure-Property Relationship) models were designed using the Gutman index, the product connectivity Banhatti index, the Variance of degree index, and the Sigma index to predict the thermodynamic properties of monocarboxylic acids. The relationship analyses between the thermodynamic properties and the topological indices were done by using the curvilinear regression method. It is used with the linear, quadratic and cubic equations of the curvilinear regression model. These regression models were then compared.

Quantitative Structure-Property Relationship Analysis Against Critical Micelle Concentration of Sulfonate-Based Surfactant Based on Semiempirical Zindo/1 Calculation

Development of anionic surfactant compound isvery important because the anionic surfactant class iswidely used in people's lives. For instance,anionic surfactantsare used as food additives and detergents. The novelcompound of sulfonate-basedsurfactantor proposed compound has predictedthe CriticalMicelle Concentration(CMC) value of experiment. Quantitative Structure-Property Relationship (QSPR)analysisbased on semiempiricalZINDO/1 calculationwas conducted to obtain QSPR equation. Theoretical predictorsor independent variable which have an influence on the value of CMC are used to construct QSPR equation. The theoretical predictors areclassified intopredictor of electronic properties, solubility and steric. A total of 108experimentalCMCbelongs to sulfonate-basedsurfactant are calculated their theoretical predictors and analyzed by multiple linear regression. The QSPR equationwhich is obtainedfromthis study contains theimportant theoretical predictors.They are solubility properties, molecular weight, molecular size and net charge of carbon atomin thepolar partof sulfonate-based surfactant. This QSPR equation couldbe used to predict the CMC value of the novelsulfonate-based surfactant.

Proton transfer reaction confined within carbon nanotubes: Density functional theory and quantitative structure–property relationship analysis

In this work, we focus our attention on chemical reactions confined within carbon nanotubes. As a result of the confinement within carbon nanotubes, novel physical and chemical properties are found for the confined materials. We consider the feasibility of proton transfer inside carbon nanotubes. To do that, we have chosen formamide as the simplest real model for exhibiting the tautomerization in DNA. We have used the quantitative structure–property relationship method, based on geometry optimization and quantum chemical structural descriptors, to illustrate the potential of using the confined space inside carbon nanotubes, which will provide comprehensive information about carbon nanotubes. All calculations have been carried out using density functional theory quantum calculations with the B3LYP functional. The geometries optimized by the Gaussian program were transferred to the computer software DRAGON to calculate pertinent descriptors that could be used in the quantitative structure–property relationship model.

Removal of pharmaceuticals and personal care products by two-stage biofiltration for drinking water treatment

Contamination of drinking water with pharmaceuticals and personal care products (PPCPs) is an issue of health concerns. To effectively control the level of PPCPs in drinking water, a pilot study employing two parallel trains of two-stage biofiltration, i.e., a sand/anthracite (SA) biofilter coupled with a biologically-active granular activated carbon (GAC) post-filter contactor, was conducted as a post-treatment after coagulation in a drinking water treatment plant. Results showed the biofiltration process could effectively remove PPCPs with an average removal of 53.4%, where the GAC contactor played the dominant role to remove 48.1% of the total PPCPs. The molecular properties determined the removability of individual PPCPs, i.e., smaller molecules with simpler structure connectivity were more likely to be removed. Based on the quantitative structure-property relationships (QSPRs) analysis, a simple regression model was proposed to predict the removability of each PPCP across the biofiltration process. The drinking water equivalent level (DWEL) quotient method was developed to assess the health risks of detected PPCPs in water samples. The biofiltration process showed efficient capacity to reduce the health risks of PPCPs with an average removal of 79%, and the PPCPs in the effluents generally would not pose adverse health effects. Pearson correlation analysis explored the possible role of nitrogenous PPCPs (N-PPCPs) as the precursors of nitrogenous disinfection byproducts (N-DBPs) in drinking waters. Aromatic nitrogen in PPCPs was found to be a significant descriptor for the formation potential of trichloroacetonitrile (TCAN). In addition, it was found that pre-filter chlorination could slightly improve the biofiltration of PPCPs.

Quantitative Structure-Property Relationship (QSPR) Analysis of ZrO2-Containing Soda-Lime Borosilicate Glasses.

Quantitative structure-property relationship (QSPR) analysis is a promising approach to correlate structural features with properties of glass materials that lack long-range order and usually have complex structures. By using carefully chosen descriptors based on structural models generated from molecular dynamics (MD) simulations, correlations with properties and insights on glass behaviors can be obtained. Zirconia can significantly alter glass properties including chemical durability, even in a small amount, and hence plays an important role in vitrification of nuclear waste where long-term chemical durability is desired. In this study, borosilicate glasses with the composition of xZrO2-(61 - x)SiO2-17B2O3-18Na2O-4CaO with x = 0, 1, 2, 4, 6, and 8 were simulated using classical MD simulations with the recently developed composition-dependent potentials. Short-range (e.g., bond distance and coordination numbers) and medium-range (e.g., Qn distribution, network connectivity, and ring-size distribution) structural features altered by ZrO2 were obtained and analyzed. The use of a descriptor ( Fnet descriptor) that combines short-range structural characteristics, from MD simulations, and the cation-oxygen single bond strength was found to provide excellent linear correlations with the density and initial dissolution rate of these glasses. The results show that by combining MD simulations and QSPR analysis the composition and structural effect on the properties of complex multicomponent glasses can be elucidated, thus suggesting that this is a promising approach for future glass research and new composition design.

Predicting melting points of hydrocarbons by the Graovac-Pisanski index

ABSTRACTTheoretical molecular descriptors alias topological indices are a convenient means for expressing in a numerical form the chemical structure encoded in a molecular graph. The structure descriptors derived from molecular graphs are widely used in Quantitative Structure-Property Relationship (QSPR) and Quantitative Structure-Activity Relationship (QSAR). In this paper, we are interested in the Graovac-Pisanski index (also called modified Wiener index) introduced in 1991 by Graovac and Pisanski, which encounters beside the distances in a molecular graph also its symmetries. In the QSPR analysis we first calculate the Graovac-Pisanski index for different families of hydrocarbon molecules using a simple program and then we show a correlation with the melting points of considered molecules. We show that the melting points of the alkane series can be very effectively predicted by the Graovac-Pisanski index and for the rest of considered molecules (PAH’s and octane isomers) the regression models are different, but we establish some correlation with the melting points for them as well.

Theoretical and Computational Investigations into Lipid Bilayer Permeation of Drugs

Transport phenomena across biomembranes are crucial processes in cellular biology, as well as in many medical, pharmaceutical, and environmental technologies. Therefore, determination of drug absorption is an important component of the drug discovery and development process in that it plays a key role in the decision to promote drug candidates to clinical trials. The path of a drug from the site of administration to its target cells or compartments implies the crossing of several semipermeable cell membranes, making it relevant to be able to predict whether and to what extent a molecule can pass through them. Lipid bilayers are an effective barrier to passive diffusion of ions and some hydrophilic small molecules, but many others can permeate bilayers through passive diffusion at rates that depend on bilayer composition and properties of the permeating solute. In this study, we consider the membrane deformation energy as the dominant factor in crossing the membrane into cells, as measured by in vitro cell-based experiments. We have investigated a new approach using the deformation free energy of a lipid bilayer based on the principle of a continuum theory. To gain atomistic insight into the passive permeability process, we have used physics-based methods, namely molecular dynamics simulations combined with the inhomogeneous solubility-diffusion model. The estimated permeabilities from our method are compared with other popular methods such as Quantitative Structure-Property Relationship (QSAR) analysis.

Prediction of High Performance Liquid Chromatography Retention Time for Some Organic Compounds Based on Ab initio QSPR Study

Analysis of the quantitative relationship between structure and characteristics of 18 polyaromatic hydrocarbons has been done by quantum Ab initio Quantitative Structure-Property Relationship (QSPR) study at Hartree Fock level of theory. Moment dipole was used as the quantum chemical descriptors, whereas molecular weight and number of rings were applied for constitutional descriptor and the valence connectivity index as steric descriptors. The compound’s electronic structure was studied by molecular modeling and retention time (Tr) data were obtained from the literature. Multi-linear regression analysis has been performed by randomly splitting the initial data set into on fitting data set and a test data set. The best result provided by QSPR analysis is the following model equation:
 log tR = 1.276 + 0.016MW+0.323Rc-0.423χ1-0.147χ2 with n =18 r=0.917 r2 =0.841 SE=0.182 Fcalc/Ftable = 5.408.
 The retention times of PAH compounds with two and three rings were successfully predicted by QSPR models.

Different encoding alternatives for the prediction of halogenated polymers glass transition temperature by quantitative structure–property relationships

ABSTRACTThe glass transition temperature, Tg, is one of the most important properties of amorphous polymers. The ability to predict the Tg value of a polymer preceding its synthesis is of enormous value. For this reason it is of great value to perform a predictive quantitative structure–property relationships analysis of Tg, in this case a new set of halogenated polymers was used for this purpose. In addition, to corroborate our previous findings, the best way to encode the polymers structure for this type of studies was further tested finding that the optimal option is once more to use three monomeric units. The best linear model constructed from 153 molecular structures incorporated seven molecular descriptors and showed excellent predictive ability. Furthermore, the method showed to be very simple and straightforward for the prediction of Tg since three-dimensional descriptors are not required.

In silico designing of power conversion efficient organic lead dyes for solar cells using todays innovative approaches to assure renewable energy for future

Advances in solar cell technology require designing of new organic dye sensitizers for dye-sensitized solar cells with high power conversion efficiency to circumvent the disadvantages of silicon-based solar cells. In silico studies including quantitative structure-property relationship analysis combined with quantum chemical analysis were employed to understand the primary electron transfer mechanism and photo-physical properties of 273 arylamine organic dyes from 11 diverse chemical families explicit to iodine electrolyte. The direct quantitative structure-property relationship models enable identification of the essential electronic and structural attributes necessary for quantifying the molecular prerequisites of 11 classes of arylamine organic dyes, responsible for high power conversion efficiency of dye-sensitized solar cells. Tetrahydroquinoline, N,N′-dialkylaniline and indoline have been least explored classes under arylamine organic dyes for dye-sensitized solar cells. Therefore, the identified properties from the corresponding quantitative structure-property relationship models of the mentioned classes were employed in designing of “lead dyes”. Followed by, a series of electrochemical and photo-physical parameters were computed for designed dyes to check the required variables for electron flow of dye-sensitized solar cells. The combined computational techniques yielded seven promising lead dyes each for all three chemical classes considered. Significant (130, 183, and 46%) increment in predicted %power conversion efficiency was observed comparing with the existing dye with highest experimental %power conversion efficiency value for tetrahydroquinoline, N,N′-dialkylaniline and indoline, respectively maintaining required electrochemical parameters.

Modeling and prediction of octanol/water partition coefficient of pesticides using QSPR methods

PurposeThe purpose of this paper is to predict the octanol/water partition coefficient (Kow) of 43 organophosphorous compounds.Design/methodology/approachA quantitative structure-property relationship analysis was performed on a series of 43 pesticides using multiple linear regression and support vector machines methods, which correlate the octanol-water partition coefficient (Kow) values of these chemicals to their structural descriptors. At first, the data set was randomly separated into a training set (34 chemicals) and a test set (nine chemicals) for statistical external validation.FindingsModels with three descriptors were developed using theoretical descriptors as independent variables derived from Dragon software while applying genetic algorithm-variable subset selection procedure.Originality/valueThe robustness and the predictive performance of the proposed linear model were verified using both internal and external statistical validation. One influential point which reinforces the model and an outlier were highlighted.