Monte Carlo based QSGFEAR: prediction of Gibb's free energy of activation at different temperatures using SMILES based descriptors. (28th September 2022)

• Record Type: Journal Article
• Authors: Singh, Rahul; Kumar, Parvin; Devi, Meena; Lal, Sohan; Kumar, Ashwani; Sindhu, Jayant; Toropova, Alla P.; Toropov, Andrey A.; Singh, Devender
• Is Part Of: New journal of chemistry. Volume 46:Number 39(2022)
• Page Start: 19062
• Page End: 19072
• Publication Date: 2022-09-28
• Subjects: Chemistry -- Periodicals; Chimie -- Périodiques; 540
• DOI: 10.1039/d2nj03515d
• Languages: English
• ISSNs: 1144-0546

Abstract

Abstract : Monte Carlo optimization based QSGFEAR model development using CII results in the formation of more reliable, robust and predictive models. Abstract : In the present manuscript, we have endeavored to construct quantitative structure Gibb's free energy of activation relationship (QSGFEAR) models having general applicability and thorough validation. The experimental data of Gibb's free energy of activation (Δ G ‡ ) at seven different temperatures was employed as the endpoint and the descriptor of correlation weight (DCW) was computed from the SMILES notation of the compounds. The QSGFEAR models were validated with a new statistical parameter called correlation intensity index (CII). A total of eight models were formed from the data set of experimentally determined Δ G ‡ values, four using target function TF1 ( W CII = 0.0) and four using target function TF2 ( W CII = 0.3). It was found that the models built by applying CII were more predictive, robust and consistent than those without CII. All the developed models were found to be effective for predicting Δ G ‡ values reliably and consistently. The leading model was the one developed from split 3 using TF2 with R Val 2 = 0.9108. The mechanistic interpretation was done with the help of split 3 and the SMILES attributes responsible for the increase and decrease of Δ G ‡ value were identified.