F. Cisternino

This study uses 3D diffusion autoencoders to derive latent cardiac MRI phenotypes, finds they share a genetic basis with cardiometabolic traits, and shows predictive power for disease risk.

The paper introduces TGlow, a high-content imaging platform for single-cell phenotyping of human lymphocytes, revealing distinct morphological states and drug responses.

This paper uses self-supervised learning to analyze histology images and predict spatial RNA expression across 23 human tissues, revealing population variability and its impact on gene expression.