QSAR / QSPR Modeling

Tools & Technologies - in silico Drug Discovery

Quantitative structure-activity relationships (QSAR) and structure-property relationships (QSPR) modeling are critical in silico tools for drug discovery project teams. These approaches help build predictive models for to guide teams during forecasting affinity to targets, as well as anti-targets and help drive effective use of medicinal chemistry resources. It is also useful for understanding the structural drivers for important molecular properties such as solubility, cell penetration, and blood-brain barrier permeability, to name a few. We use a wide range of molecular descriptors (2D, 2.5D, and 3D) and statistical and machine learning approaches such as Bayesian modeling, principal component analysis, partial least squares, genetic-partial least squares, random forest, support vector machines, ant-colony optimization, and Kohonen neural nets to build predictive models to guide teams during compound optimization.