Home / Research Areas / Precision & Genomic Medicine
Same Diagnosis, Different Biology — Find the Treatment That Fits the Patient
Most treatment guidelines are built around how the average patient responds. But people differ — genetically, biologically, in how they process drugs — and those differences can be the gap between a treatment working and a treatment failing. This pathway trains you to find the patient subgroups hiding inside a single diagnosis.
Finding the Subgroup Inside the Diagnosis
Network and systems approaches help find subgroups of patients within a single disease who share an underlying molecular signature — even when they look clinically identical on the surface.
— Cluster patients into subgroups using genetic, molecular, or clinical similarity
— Study pharmacogenomics — how genetic variants affect how patients metabolize and respond to specific drugs
— Evaluate candidate biomarkers for real clinical usefulness, not just statistical significance
— Work with large biobank datasets linking genetics to real health outcomes over time
Closing the Gap Between "Interesting" and "Useful"
Precision medicine is already changing clinical practice — genetic testing before certain prescriptions, molecular subtyping before choosing cancer treatment, risk scores combining dozens of genetic variants. The gap between "we found something statistically interesting" and "this changes what a doctor does" is where most precision medicine research effort actually goes — and it's exactly what this area trains students to evaluate.
What a Summer Project Looks Like
Precision & Genomic Medicine
Do Genetic Subtypes Predict Treatment Response? A Patient Stratification Study
Using a publicly available biobank or cohort dataset, a student identifies a disease with known clinical heterogeneity. They apply clustering and dimensionality-reduction methods to multi-omics or genetic data to identify distinct patient subgroups, then investigate whether those subgroups correspond to meaningfully different outcomes or treatment responses. The project includes a critical evaluation step: is this subgrouping statistically real, and would it actually hold up in a new group of patients? Mentored by a clinical genomics researcher or biostatistician, the student produces a written analysis and a visualization of the patient subgroups discovered.
Clustering & Dimensionality ReductionPharmacogenomicsBiomarker EvaluationBiobank-Scale Data
Precision & Genomic Medicine
Do Genetic Subtypes Predict Treatment Response? A Patient Stratification Study
Using a publicly available biobank or cohort dataset, a student identifies a disease with known clinical heterogeneity. They apply clustering and dimensionality-reduction methods to multi-omics or genetic data to identify distinct patient subgroups, then investigate whether those subgroups correspond to meaningfully different outcomes or treatment responses. The project includes a critical evaluation step: is this subgrouping statistically real, and would it actually hold up in a new group of patients? Mentored by a clinical genomics researcher or biostatistician, the student produces a written analysis and a visualization of the patient subgroups discovered.
Clustering & Dimensionality Reduction
Pharmacogenomics
Biomarker Evaluation
Biobank-Scale Data
How Each Level Gets There
Every track takes a shared Fall foundations course, then this area-specific Spring course, then a live mentored summer project.
High School
HS-S-PM
Undergraduate
UG-S-PM
Medical Student
MS-S-PM
Clinician
CL-S-PM
Not sure this is the right area for you? Explore all six research areas →