SiliS-PTOXRA, acronym of DRY-in Silico Screening and Prediction of Toxic Hazard and Risk Assessment, was developed for the prediction of toxicological endpoints. This software was developed in Java 1.8; thus, it is fully cross-platform. For the development of this program, the CDK (v1.4.19), QuBiLS-MAS, and AMBIT (v0.0.4) libraries were used for the handling of chemical structures, computation of molecular descriptors, and computation of applicability domain (AD). All available models in this software were built with the WEKA (v3.8.0) software.

Currently, the available QSAR models to predict toxicological endpoints are:

  • Acute Oral Toxicity (AOT): Regression models based on topological (2D) QuBiLS-MAS molecular descriptors
  • Drug-induced liver injury (DILI): Binary classification models based on topographical (2.5D) QuBiLS-MAS molecular descriptors
  • Acute Toxicity towards Zebrafish Embryo: multiple linear regression models based on topological (2D) QuBiLS-MAS molecular descriptors
  • Acute Toxicity towards Freshwater Crustacean Thamnocephalus Platyurus: multiple linear regression models based on topological (2D) QuBiLS-MAS molecular descriptors
  • Inhalation Toxicity using No Observed Adverse Effect Concentration (NOAEC): multiple linear regression models based on topological (2D) QuBiLS-MAS molecular descriptors
Graphical User Interface (GUI)