Atlas

Inspiration

Our inspiration for Atlas comes from the importance of the patient perspective in rare disease research. In many ultra-rare conditions, traditional clinical trial design fails because patient populations are extremely small and heterogeneous. Patients and advocacy organizations often collect valuable longitudinal data such as symptoms, biomarker measurements, and treatment experiences. However these datasets are rarely structured in a way that allows researchers to translate them into meaningful clinical trial endpoints. We wanted to build a tool that bridges that gap. Atlas centers the patient voice by helping researchers identify biomarkers and endpoints that reflect real patient experiences while still being measurable and scientifically rigorous. By empowering patient communities to contribute structured data insights, Atlas helps ensure that future trials pursue outcomes that truly matter to patients.

How we built it

We built Atlas using Posit Cloud with the Shiny framework, allowing us to create an interactive web-based analytics platform. The backend processes patient datasets uploaded as Excel files and cleans and standardizes the data using R packages such as dplyr, tidyr, and readxl. Atlas then analyzes genotype–phenotype relationships and calculates biomarker detectability by measuring within-patient variability and integrating feasibility metadata such as cost, measurement setting, and confounder risk. The application includes simulation tools that model how potential biomarkers would perform as clinical trial endpoints in small cohorts. The Shiny interface allows users to explore patient-level data, visualize biomarker trends, and run simulations that estimate statistical power in ultra-small clinical populations.

Challenges we ran into

One of our biggest challenges was learning how to code and design analytical workflows with limited prior programming experience. Building Atlas required us to rapidly learn how to structure data pipelines, implement statistical simulations, and construct a responsive Shiny interface. Debugging and integrating multiple R packages also required trial and error. Another unexpected challenge was preparing the project for submission on Devpost while simultaneously refining the application itself. The process pushed us to quickly develop new technical skills and deepen our understanding of how computational tools can support rare disease research.

Accomplishments that we're proud of

We are proud of building a functional tool that helps researchers analyze extremely small patient datasets, a persistent challenge in rare disease research. It helps integrate the patient experience along with actionable data to create viable biomarker candidates. Atlas demonstrates that meaningful insights can be extracted even from limited patient cohorts by combining patient-reported outcomes, biomarker data, and statistical simulation. Most importantly, we kept the patient experience at the center of the project. The platform is designed to use patient data and turn it into endpoints that are both clinically useful and measurable in trials. We are proud to create a system that empowers patient advocacy organizations and researchers to collaborate more effectively.

What we learned

Through this project, we learned how to adapt data analysis methods typically used for large biomedical datasets to the unique constraints of rare diseases, where sample sizes are extremely small. We also gained experience building end-to-end analytical tools, from data ingestion and cleaning to visualization and statistical modeling. Perhaps most importantly, we learned how computational tools can help patient communities organize their data in ways that make it more actionable for researchers and clinicians. The project reinforced how interdisciplinary collaboration between patients, scientists, and technologists can accelerate progress in rare disease research.

What's next for Atlas

Our initial audience for Atlas is ultra-rare disease patient advocacy organizations, many of which already collect small but valuable datasets about their communities. By providing these groups with an accessible analytical platform, Atlas allows them to organize longitudinal biomarker and symptom data in ways that support research and clinical trial development. In the future, key opinion leaders in rare disease medicine could use Atlas to track biomarker trends across patient registries and identify promising endpoints for therapeutic studies. As the platform grows, biotech and pharmaceutical companies could also leverage Atlas to explore biomarker signals within extremely small patient populations. Ultimately, the long-term goal of Atlas is to help identify reliable biomarkers that justify and enable the development of clinical trials for diseases that currently lack viable therapeutic pathways.