Precision Medicine: Personalized Healthcare in the Genomic Era

From One-Size-Fits-All to Precision

Traditional medicine: same treatment for all patients with same diagnosis. Problem: drugs work in only 30-60% of patients (Spear et al., 2001). Precision medicine: tailoring prevention, diagnosis, and treatment based on individual genetic, environmental, and lifestyle factors. Human Genome Project (2003): first complete human genome, $2.7 billion. Today: whole genome sequencing ~$200, turnaround 24-48 hours. GWAS (Genome-Wide Association Studies): identified thousands of genetic variants associated with diseases. Polygenic risk scores: combining many variants to predict disease risk.

Pharmacogenomics

Pharmacogenomics: how genetic variation affects drug response. CYP450 enzymes: metabolize ~75% of drugs. CYP2D6 polymorphisms: poor metabolizers (7% of Caucasians) vs. ultra-rapid metabolizers — codeine toxicity vs. inefficacy. CYP2C19: clopidogrel activation — FDA black box warning for poor metabolizers. DPYD: fluoropyrimidine toxicity (5-FU), 3-8% carry reduced-function variants. HLA-B*5701: abacavir hypersensitivity (HIV) — mandatory testing before prescribing. Warfarin: VKORC1 and CYP2C9 variants affect dosing 2-3×. CPIC guidelines: Clinical Pharmacogenetics Implementation Consortium — evidence-based genotype-drug guidelines for 40+ drugs.

Cancer Precision Oncology

Molecular tumor profiling: next-generation sequencing of tumor DNA. Actionable mutations: EGFR (lung cancer — erlotinib, osimertinib), HER2 (breast — trastuzumab), BRAF V600E (melanoma — vemurafenib), ALK fusion (lung — crizotinib). TMB (Tumor Mutation Burden): high TMB → better response to immunotherapy (pembrolizumab). MSI-H (Microsatellite Instability-High): tissue-agnostic FDA approval for pembrolizumab. Liquid biopsy: circulating tumor DNA (ctDNA) from blood — non-invasive monitoring. Guardant360, FoundationOne Liquid CDx — FDA-approved panels. ctDNA for minimal residual disease (MRD) detection after surgery. CAR-T therapy: genetically modified T-cells targeting CD19 (Kymriah, Yescarta for lymphoma).

Rare Disease Genomics

Rare diseases: ~7,000 known, affecting 300 million people worldwide. 80% have genetic origin. Diagnostic odyssey: average 5-7 years to diagnosis. Whole exome sequencing (WES): analyzes protein-coding regions (1.5% of genome, ~85% of disease-causing mutations). Diagnostic yield: 25-40%. Whole genome sequencing (WGS): captures non-coding variants, structural variants. Undiagnosed Diseases Program (NIH): solving the unsolvable. Gene therapy successes: Luxturna (RPE65 retinal dystrophy), Zolgensma (SMA, AAV-delivered SMN1 — most expensive drug: $2.1M). Base editing: VERVE-101 for familial hypercholesterolemia (PCSK9 editing in liver).

AI in Precision Medicine

AI for drug discovery: target identification, lead optimization, clinical trial design. DeepVariant (Google): CNNs for genomic variant calling — accuracy exceeding traditional tools. PathAI: AI-assisted pathology for cancer grading. Tempus: genomic + clinical data integration for oncology treatment selection. Foundation models for genomics: Evo (Arc Institute), Nucleotide Transformer — predicting variant effects from DNA sequence. Polygenic risk scores + AI: predicting disease 5-10 years before onset. Challenges: data privacy (HIPAA, GDPR), algorithmic bias in underrepresented populations, clinical validation requirements, integration into electronic health records, reimbursement models.

Try it live

Everything above runs in your browser — open Cardiac Action Potential and change the parameters while it is running. Nothing is installed, nothing is uploaded, the whole model lives in one tab.