Skip to Main content Skip to Navigation

Modeling and visualization of complex chemical data using local descriptors

Abstract : This work describes original approaches for predictive chemoinformatics modeling of molecular interactions and reactions as a function of the structures of interacting partners and of the chemical environment (experimental conditions). Chemical structures have been encoded by local ISIDA MA-based or CGR-based descriptors, specifically targeting the active centers and their closest environment. The local descriptors have been combined with the specific parameters of experimental conditions, thereby encoding a particular chemical object. The methodology has been successfully applied for QSPR modeling of thermodynamic and kinetic parameters of intermolecular interactions (halogen and hydrogen bonds), tautomeric equilibria and chemical reactions (cycloaddition and SN1). GTM method has been applied for the first time for QSPR modeling and visualization of mixed chemical data. This method successfully separates data clusters on account of both chemical structures and experimental conditions.
Document type :
Complete list of metadata
Contributor : Abes Star :  Contact
Submitted on : Tuesday, December 4, 2018 - 10:33:11 AM
Last modification on : Wednesday, October 14, 2020 - 4:08:58 AM


Version validated by the jury (STAR)


  • HAL Id : tel-01943762, version 1



Marta Glavatskikh. Modeling and visualization of complex chemical data using local descriptors. Cheminformatics. Université de Strasbourg; Kazanskij gosudarstvennyj universitet im. V. I. Ulʹânova (Kazanʹ), 2018. English. ⟨NNT : 2018STRAF008⟩. ⟨tel-01943762⟩



Record views


Files downloads