Biomedical and Health Sciences

Analyze TCGA splicing data to find tumor-specific cryptic exons, rank neoantigen candidates, and practice real cancer genomics analysis.

Learn how to cluster sepsis ICU stays with temporal embeddings and test whether fluid response differs across patient phenotypes.

Build a computational workflow to design miniprotein binders and score them with structure prediction, confidence, and energy checks.

Map asthma deserts with public data, then test how pollution and clinic access combine to predict pediatric ER visits by ZIP code.

Build a model that predicts acute kidney injury before diagnosis, and learn how transformers, lab trends, and SHAP make ICU data readable.

Build a satiety model from public trials, then test how protein density predicts hunger, fullness, and intake in a 14-day self-study.

Build an offline pill recognition app and test how lighting, background clutter, and phone hardware change identification accuracy.

Use public scRNA-seq data to map pancreatic cancer fibroblast states and build a short marker gene set with clear classification power.

Map Triphala compounds to disease targets, build interaction networks, and use docking ideas to test how plant molecules may affect metabolic syndrome.

Measure how blood pressure changes by setting, then fit a Bayesian model to separate context effects from a teen’s true baseline.