Data Driven Drug Discovery

This webinar provides an introduction to quantitative structure-activity relationship (QSAR) modeling. QSAR modeling is useful for understanding relationships between chemical structure and the biological or pharmacological action of compounds. Inevitably, many such relationships are multivariate by nature, where groups of key chemical variables jointly influence biological behavior. Multivariate methods are ideal tools for understanding these complex relationships, and for directing research towards compounds with enhanced biological performance.

In this webinar we show the utility of the data analytics toolbox of SIMCA 15 in deriving sound, transparent and interpretable QSARs. Several QSAR datasets from the fields of drug design, pesticide research, and environmental toxicology and chemistry are worked out in detail, showing the benefits of PCA, OPLS and multivariate design. PCA is useful when overviewing a dataset and exploring relationships among compounds and among variables. OPLS is the regression extension of PCA and is used for establishing QSARs. Multivariate design is essential for selecting an informative training set of compounds for QSAR calibration. Finally, SIMCA´s ability to interpret SMILES code and visualizing molecular structure is exemplified.