MAGELLAN

2026 Therapeutic Targets Winner

Magellan is an AI-driven compatibility engine that analyzes public biomedical data on existing drugs and rare diseases to identify repurposing opportunities. By comparing disease biology and drug mechanisms across biological pathways, Magellan ranks potential drug–disease matches, highlighting the most promising indications for existing drugs and providing quantitative compatibility scores with supporting evidence to accelerate rare disease therapeutic discovery.

PROJECT SUMMARY

Rare disease drug development remains one of the most difficult challenges in medicine. While over 10,000 rare diseases have been identified, only 5% of them have FDA-approved treatments. At the same time, pharmaceutical companies possess thousands of shelved or deprioritized compounds, as well as a large universe of generic or patent-expiring drugs that may have therapeutic potential, but have never been systematically evaluated for rare disease indications.

Magellan is a drug–disease compatibility engine designed to identify potential rare disease treatments through drug repurposing. Users can input an existing drug, and Magellan ingests relevant clinical, mechanistic, and safety data from multiple biomedical sources, normalizes this information into a structured mechanistic framework, and ranks candidate drug–disease pairs. Each result includes mechanistic nodes involved, evidence links, safety flags, and an explicit uncertainty score to provide transparent, explainable reasoning. The platform is built with a Python 3.11 backend using FastAPI, Celery, and SQLAlchemy in a Docker environment, supported by PostgreSQL and Redis, and integrates public biomedical APIs such as PubChem, ChEMBL, ClinicalTrials.gov, PubMed/NCBI, ClinVar, openFDA, and Orphanet to compute mechanism-level drug–disease scores and supporting evidence.

Magellan is designed to support researchers, biotechnology companies, and rare disease foundations seeking to identify new therapeutic strategies using existing drugs. By prioritizing repurposing opportunities with mechanistic justification and transparent evidence, the platform aims to reduce development risk and accelerate the discovery of treatments for underserved patient populations. Future development could include integrating additional proprietary datasets, improving predictive models, validating predictions through experimental or clinical data, and expanding the platform to evaluate thousands of drugs across a broad range of rare diseases.

Ultimately, Magellan aims to transform how society thinks about abandoned or generic drugs, turning them from sunk costs into continuously mined therapeutic opportunities and helping close the gap between the thousands of known rare diseases and the small fraction that currently have approved treatments.

MEET THE TEAM

Alexander Kang
University of Pennsylvania
Undergraduate (2028)
Finance and Biology

Jishnuu Senthil Kumar
University of Pennsylvania
Undergraduate (2029)
Bioengineering and Computer Science

David Oleksy
Harvard University
Undergraduate (2028)
Biology

Shiven Sasipalli
Brown University
Undergraduate (2026)
Neuroscience