Apply

Fellows accepted into the program will gain insight into research in both academia and industry. Our prestigious fellowships are fully funded for a three-year period and provide the fellows’ salaries, fringe benefits, and a generous travel allowance. Our fellowships are open to outstanding candidates seeking career advancement in biomedical data sciences.

The Yale-BI Joint Steering Committee (JSC) has established the following data-driven research theme for the 2025 program. The steering committee will consider how well proposed research projects align with the prioritized theme when judging submissions and postdoctoral applications. Selected applications will be further reviewed for approval.

We are seeking research proposals supported by the Yale-Boehringer biomedical data science fellowship program in 2025. Successful candidates will be mentored by both Yale and BI mentors, leveraging the world-leading ML/AI research capabilities at Yale and the strong drug discovery expertise at Boehringer Ingelheim to develop novel methods in new target and biomarker discovery and precision patient treatment. Through novel computational and AI innovation, fellows will have the opportunity to generate new insights, impact one or more of the following chronic diseases with unmet medical needs: chronic liver or kidney diseases, fibrotic inflammatory diseases, and cancer.

• Develop a foundation model-based approach using patient-level clinical and omics data to predict patients who will respond to a given therapeutic intervention (e.g., a VEGF inhibitor) in a few diseases (e.g., several different cancer types).
• Leverage large foundation model architecture to learn general biological relationships (e.g., organ-specific), and use associated embeddings in a transfer-learning framework to be able to predict patient responses in multiple diseases, preferably in few-shot learning settings.
• Model architecture may enable one or more drug discovery relevant applications. This can include but is not limited to: discovery of novel therapeutic interventions based on learnings with perturbation screening data, identification of patient populations for new therapeutic concepts in pre-clinical stages or patient selection in clinical trial design, or identification or selection of patient stratification biomarkers.

• Context-aware large language models to create large-scale, interpretable pathway representations of human biology and diseases. Able to render pathways or networks in different contexts, e.g., different diseases, disease stages, cell types, or organs.
• Create an atlas of pathways that leverage large language models to mine literature data, leveraging large-scale publicly available experimental data where possible (e.g.; human cell atlas, CRISPR screens). Can be extended to include inter-cellular networks such as ligand-receptor interactions revealed from spatial omics studies.
• Applications can be either broad rendering of human disease biology, or deep representations that focus on specific human organs or disease patho-mechanisms.

• Leverage foundation models to build novel algorithms to predict functional (ie molecular, pathway, or phenotypic) effects of either SNP or gene-level perturbations.
• Leverage large scale biobanks (e.g., All of Us) or other clinical genetics datasets to validate AI-based inferences to identify causal gene-disease links.
• Application to patient-level data and enabling improved understanding of clinical efficacy and/or toxicity of therapeutic perturbations.