Among the new families of effective anticancer drugs, the natural product paclitaxel (Taxol/Bristol-Myers-Squibb) and its semisynthetic derivative docetaxel (Taxotere/Rhone-Poulenc Rorer) are probably the most promising agents under investigation. Surprisingly considering their importance no detailed quantum mechanical studies have been carried out for these drugs. In this work we report the first structure--activity relationship (SAR) studies for 20 taxoid structures using molecular descriptors from all-electron quantum methods. The used methods were the pattern-recognition Principal Component Analysis (PCA), Hierarchical Clustering Analysis (HCA), and the recently developed Electronic Indices Methodology (EIM). The combined use of EIM with PCA/HCA methodologies was able to correctly classify active and inactive taxoids with 100% of accuracy using only a few universal quantum molecular descriptors. It was possible to identify the electronic features defining active molecules. This information can be used to select and design new active compounds. The combined use of EIM with PCA/HCA can be a new and very efficient tool in the field of computer assisted drug design.43699-706