Predictive modelling of the LD<sub>50</sub> activities of coumarin derivatives using neural statistical approaches: Electronic descriptor-based DFT

Lakhlifi

et al. 2017

Self Cite

The ACD/ChemSketch, MarvinSketch, and ChemOffice programmes were used to calculate several molecular descriptors of 51 odorant molecules (15 alcohols, 11 aldehydes, 9 ketones and 16 esters). The best descriptors were selected to establish the Quantitative Structure-Property Relationship (QSPR) of the retention/release property of odorant molecules in pectin gels using Principal Components Analysis (PCA), Multiple Linear Regression (MLR), Multiple Non-linear Regression (MNLR) and Artificial Neural Network (ANN) methods We propose a quantitative model based on these analyses. PCA has been used to select descriptors that exhibit high correlation with the retention/release property. The MLR method yielded correlation coefficients of 0.960 and 0.958 for PG-0.4 (pectin concentration: 0.4% w/w) and PG-0.8 (pectin concentration: 0.8% w/w) media, respectively. Internal and external validations were used to determine the statistical quality of the QSPR of the two MLR models. The MNLR method, considering the relevant descriptors obtained from the MLR, yielded correlation coefficients of 0.978 and 0.975 for PG-0.4 and PG-0.8 media, respectively. The applicability domain of MLR models was investigated using simple and leverage approaches to detect outliers and outside compounds. The effects of different descriptors on the retention/release property are described, and these descriptors were used to study and design new compounds with higher and lower values of the property than the existing ones.

Section: Multiple Linear Regression (Mlr)mentioning

confidence: 56%

QSPR study of the retention/release property of odorant molecules in pectin gels using statistical methods

Belhassan

Lakhlifi

et al. 2017

Self Cite

“…Three components constitute a neural network: the processing elements or nodes, the topology of the connections between the nodes, and the learning rule by which new information is encoded in the network. Although there many different ANN models, the most frequently used type of ANN in QSAR is the three-layered feed forward network [17]. In this type of network, the neurons are arranged in layers as an input layer, one hidden layer and an output layer.…”

Section: Discussionmentioning

confidence: 99%

QSAR analysis of the toxicity of phenols and thiophenols using MLR and ANN

Ghamali

Ousaa

et al. 2017

Self Cite

This study gives a quantitative structure-activity relationship (QSAR) analysis of toxicity of phenols and thiophenols to Photobacterium phosphoreum, which is an important indicator for water quality. The chemical structures of 51 phenols and thiophenols have been characterized by electronic and physic-chemical descriptors. The present study was performed using principal components analysis (PCA), multiple regression analysis (MLR) and artificial neural network (ANN). The quantitative model was accordingly proposed and the toxicity of the compounds was interpreted based on the multivariate statistical analysis. This study shows that the results obtained by MLR were suitable and have served to predict toxicity, but compared to the results of the ANN model, we conclude that the prediction achieved by the latter is more effective and better than MLR model. The statistical results of the predictive performance indicate that the ANN model is statistically significant than the previous reported model based on CART-LS-SVR. Following to the obtained results, our proposed model may be useful to predict of toxicity and risk assessment of chemicals.

“…Three components constitute a neural network: the processing elements or nodes, the topology of the connections among the nodes, and the learning rule by which new information is encoded in the network. Although there many different ANN models, the most frequently used type of ANN in QSAR is the three-layered feed-forward network [31]. In this type of network, the neurons are arranged in layers as an input layer, a hidden layer and an output layer.…”

Section: Discussionmentioning

confidence: 99%

Combining DFT and QSAR computation to predict the interaction of flavonoids with the GABA (A) receptor using electronic and topological descriptors

Ghamali

Aouidate

et al. 2017

To establish a quantitative structure-activity relationship model of the binding affinity constants (−log K i) of 41 flavonoid derivatives towards the GABA (A) receptor, the DFT-B3LYP method with basis set 6-31G (d) was performed to gain insights into the chemical structure and property information for the studied compounds. The best topological and electronic descriptors were selected. This work was conducted with principal component analysis (PCA), multiple linear regression (MLR), multiple non-linear regression (MNLR) and artificial neural network (ANN). According to these analyses, we propose quantitative models and interpret the activity of the compounds based on multivariate statistical analysis. The statistical results of the MLR, MNLR and ANN indicate that the determination coefficients R 2 were 0.896, 0.925 and 0.916, respectively. The results show that the three modelling methods can predict the studied activity well and may be useful for predicting the biological activity of new compounds. The statistical results indicate that the models are statistically significant and stable with data variation in the external validation.