(Quantitative) Structure Activity Relationships (Q)SAR
QSAR models are powerful tools to predict chemically induced toxicological effects for substances with few or no experimental test data. These models can reduce the number of animal tests necessary in the assessment of chemical substances. They can increase the information for a given substance (including metabolites and degradation products), and to save time and money.
(Q)SARs are relationships between structural properties of chemical substances and another property. This other property can be a physico-chemical property or a biological activity, including the ability to cause toxic effects.
The (Q)SARs can either be statistical models established on the basis of experimental test data for a set of chemical substances with available test data, or rule-based models based on available test data and literature information on toxicity mechanisms.
Once built, a QSAR can be used to predict the modelled property for compounds that lack experimental test data, without the need for experimental testing. For such predictions to be reliable, the given compound needs to be of sufficiently similarity to the compounds used to build the QSAR, i.e. be in applicability domain of the given (Q)SAR.