Let’s share data to see whether our drinking water is clean enough
Somewhat oddly placed within the IWA Busan congress session ‘Improvement of conventional water treatment-organic matter removal’, the QSAR modelling approach was explained by the aid of an advanced oxidation process. However, the subject of the presentation fitted within the idea to remove organic matter, since the generated QSARs (Quantitative Structure Activity Relationships) could well predict the degradation of pharmaceuticals and herbicides in the UV advanced oxidation process (AOP).
Scientific researcher Bas Wols discovered that these QSARs can have high predictability, but their reliability is sensitive to a representative (compound) data set with enough data points and an accurate process model.
Today I also explained that the AOP case study revealed that some process model parameters are sensitive to the structure properties of a (e.g. pharmaceutical) compound and therefore better to predict by QSARs, while other process parameters may hardly be sensitive. The bottom line of our work: most of all, it is of utmost importance to share removal efficiency data and use rigorous data validation methods between research organizations and universities for as many compounds as possible, to get the best out of QSAR predictions and avoid expensive, duplicate or laborious laboratory or pilot scale testing.