Introduction | Prerequisites | 1. Load the data | 1.1 HPO gene-phenotype associations | 1.2 DrugBank targets | 2. Link drug targets to phenotypes | 3. Keep the informative phenotypes | 4. Group phenotypes into clinical domains | 5. Interpret the phenotypes using the direction of drug action | Bivalirudin and F2: phenotypes predict the main adverse effect | Leuprolide and the GnRH axis: phenotypes predict the therapeutic effect | Calcitriol and VDR: the direction is reversed | 6. Compare the phenotype profiles | Conclusion

Introduction | Scientific case: finding a mechanistic substitute for lepirudin | Prerequisites | 1. Parse the DrugBank records | 2. Target actions now carry the drug | Why the drug link matters | 3. Map the targets to genes | 4. Find drugs that share lepirudin's mechanism | 5. Compare the wider interaction profiles | Summary

Introduction | 1. Loading Data | The Parsing Workflow | Loading Sample Data | 2. Analysis: The Drug Landscape | 3. Analysis: Approval Status | 4. Analysis: Molecular Mechanisms | 5. Next Steps: Integrated Pharmacovigilance

The "Hub and Spoke" Architecture | Pipeline-Ready Pharmacovigilance | New Tools for Reproducibility | A Scientific Case Study | Availability

Introduction | Prerequisites | 1. Loading and Inspecting Data | 1.1 Load DrugBank (The Hub) | 1.2 Load OnSIDES (Single-Drug Labels) | 1.3 Load TWOSIDES (Drug Interactions) | 2. Merging the Databases (Chaining) | 3. Single Drug Analysis: Leuprolide (DB00007) | 3.1 Extract Baseline Adverse Events (OnSIDES) | 4. Single Drug Analysis: Calcitriol (DB00136) | 5. Polypharmacy Analysis: Drug-Drug Interactions | 5.1 Visualization of Interaction Risks | 6. Comparative Overview | 7. Conclusion

Overview | Installation | Package Version | Naming Convention | Dataset Contents | Dataset Hierarchy