Overview

• AAnet employs deep learning to detect continuous patterns in single-cell gene expression and map tumor cell heterogeneity.
• The tool distinguished five cancer cell archetypes in breast tumors, each exhibiting unique gene expression signatures linked to proliferation, metastasis, and resistance.
• Archetype-based classification could enable oncologists to design combination therapies tailored to each cell group’s molecular vulnerabilities.
• Beyond breast cancer, the framework has potential for characterizing cellular diversity in other cancers and heterogeneity-driven diseases.
• Researchers will conduct longitudinal studies to track the evolution of these archetypes before and after treatments and advance targeted therapy strategies.