Spatial biology technologies provide complementary information about tissue anatomy but are often challenging or costly to combine experimentally. The aim of this project is to create a multi-modal generative model of spatial biology data that integrates diverse data types and can be used for cross-modality data transfer. Application areas include synthetic data generation, histology-guided gene expression imputation and super resolution in sequencing-based spatial transcriptomics, and feature imputation in high-resolution in situ data.