Flexible discovery of disease-associated tissue structures

Researchers developed VIMA, a deep learning method that identifies disease-linked tissue patterns in spatial molecular data without needing prior cell type annotations, showing strong performance across technologies and diseases.