Suitable in silico prediction

means a prediction of the hazard characteristics of an industrial chemical using a computational approach (either a quantitative prediction or a prediction indicating the absence or presence of alerting groups) for which all of the following apply: 

• the prediction is for the industrial chemical itself 
• the in silico model used for the prediction is listed in Appendix 8.2 of the Categorisation Guidelines as acceptable for the hazard characteristic 
• the industrial chemical is within the chemical space for which the in silico model gives reliable predictions (the applicability domain) - the model applicability could be in relation to any of:
  • (sub)structures present in the training set 
  • a range of molecular descriptors 
  • metabolites/degradation products 
  • mechanisms of action 
  • the guidelines, instructions and parameters for the in silico model have been followed in order to get the prediction 
  • the industrial chemical is not any of the following types of chemicals:
    • UVCB substance
    • polymer
    • surfactant
    • inorganic chemical 
    • organometallic chemical 
    • a chemical that is a solid or in a dispersion at the time of introduction, and consists of solid particles in an unbound state or as an aggregate or agglomerate, where at least 50% (by number size distribution) of the particles have at least one external dimension in the nanoscale, and is not soluble