Sylvesters-Oyim/african-pgx-pipeline

# Sovereign African PGx Pipeline > **The problem:** A doctor in Nairobi prescribes Clopidogrel (a blood thinner) after a heart procedure. The patient is also on Efavirenz for HIV. Unknown to the doctor, the patient's genetics mean their body *cannot activate* the drug — and the HIV medication makes it even worse. The patient is at serious risk of a blood clot. Today, almost no hospital in Sub-Saharan Africa can catch this before it's too late. > > **This pipeline catches it.** ## What It Does This is a clinical decision support API that reads a patient's genetic profile and tells the prescribing doctor — in plain, actionable language — whether a medication is safe, risky, or dangerous for *that specific patient*, factoring in the drug interactions that are uniquely common across East African clinical settings (HIV and TB regimens). **In plain terms, the pipeline:** 1. **Receives anonymized genetic data** from a hospital's electronic health record system — no patient names, no raw DNA, just coded gene variants 2. **Looks up how the patient's body processes the drug** using a proprietary, encrypted reference matrix built on international clinical pharmacogenomics guidelines (CPIC) 3. **Flags dangerous drug interactions** specific to the region — for example, Efavirenz (a first-line HIV drug in East Africa) blocks the same liver enzyme that activates Clopidogrel, compounding genetic risk 4. **Generates a traffic-light clinical report** (RED = stop, YELLOW = caution, GREEN = safe) as a tamper-proof PDF streamed directly back to the hospital system ### Technical Summary For developers: the API accepts star-allele diplotypes (e.g., CYP2C19 \*1/\*17), calculates activity scores from a Fernet-encrypted allele matrix, resolves CPIC phenotypes, applies regional drug-drug-gene interaction logic, and streams a SHA-256-fingerprinted PDF report. ## Supported Genes & Drugs | Gene | Drug | Guideline Basis | |------|------|----------------| | CYP2C19 | Clopidogrel | CPIC An …