SpaFactor: New Model for Gene Program Inference from Histology Images

VERA Brief

AI-generated. Grounded in the article and its cited sources.

Researchers have developed SpaFactor, a new framework for predicting gene expression from histology images. This model integrates visual data with neighborhood context to capture cellular morphology and microenvironmental diversity, aiming to improve the inference of gene expression patterns.

Key facts

• SpaFactor is a new framework for inferring gene expression patterns from histology images.
• The model uses a low-rank morphology-program-gene factorization approach.
• SpaFactor integrates visual data from tissue images, including neighborhood context.
• The framework demonstrated strong performance across five public cohorts for spatially variable genes.
• SpaFactor is designed to be efficient and computationally less intensive than previous approaches.