AI + Genomics for Cancer & Human Disease
We build interpretable AI to turn genomes into mechanisms—and mechanisms into better prevention, diagnosis, and treatment.
We build interpretable AI to turn genomes into mechanisms—and mechanisms into better prevention, diagnosis, and treatment.
AI-Driven Precision Treatment
From tumor transcriptomes, we prioritize drug candidates, combinations, and gene targets, with patient-level, explainable predictions.
2. Predictive Biology
Large language models forecast emergent gene/protein traits and interpret variants, linking sequence to mechanism and phenotype.
3. Immune Response Analysis
Single-cell and spatial data map tumor–immune circuits and microenvironment states to reveal biomarkers of response and resistance.
We pair model development with experimental validation and share models, data, and benchmarks to accelerate translation.
See the presentation here
Our latest publication in Nature Communications, in collaboration with Rafa Irizarry's Lab, introduces scDist—a principled approach to address variability in single-cell data. This rigorous tool enhances accuracy and uncovers robust perturbations in DC, pDC, and NK cells in COVID-19 and immunotherapy responses. As single-cell datasets grow, we’re excited to advance the intersection of AI, statistics, and biology.
unveils a new class of bipotent drugs for cancer therapy. This approach targets key regulators to simultaneously shut down cancer pathways and boost tumor immunity. Read this publication here. | View all our Publications.