We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
QSTR based on Monte Carlo approach using SMILES and graph features for toxicity toward Tetrahymena pyriformis.
- Authors
Rezaie-keikhaie, Nasrin; Shiri, Fereshteh; Ahmadi, Shahin; Salahinejad, Maryam
- Abstract
Tetrahymena pyriformis, due to the direct contact its cells have with the outside environment, is attractive for assessing environmental toxicant effects. We report a quantitative structure–toxicity relationship (QSTR) model on a set of 644 organic compounds based on the logarithm of 50% growth inhibitory concentration of Tetrahymena pyriformis (pIGC50). Models were implemented using the CORAL software, which did not require 3D optimization of molecules. From the SMILES format of compounds, two types of descriptors, including SMILES-based and Graph-based, were generated. The data were randomly divided into four sets: training, invisible training, calibration, and validation. Ten QSTR models have been constructed using a hybrid optimal descriptor based on the Monte Carlo algorithm. Various statistical features of constructed models depicted that all splits are robust and predictable at a high level. pIGC50 is influenced by Morgan fingerprints and fragment structure descriptors. Interestingly, the promoters of increase in Tetrahymena pyriformis toxicity are attributes related to carbon and hydrogen atoms, while those of decrease in toxicity relate to oxygen.
- Subjects
TETRAHYMENA pyriformis; SMILING; HYDROGEN atom; POISONS; ORGANIC bases; CORALS; HUMAN fingerprints
- Publication
Journal of the Iranian Chemical Society, 2023, Vol 20, Issue 10, p2609
- ISSN
1735-207X
- Publication type
Article
- DOI
10.1007/s13738-023-02859-x