Recent advances in imaging and molecular profiling technologies have enabled the acquisition of data at subcellular resolution. However, integrating such data across modalities, spatial scales, ...
and acquisition protocols remains a major challenge due to differences in resolution, feature types, and physical coverage.
In this talk, I will present a suite of technologies we have developed for cross-modality 3D mapping, enabling non-rigid alignment of transcriptomic data at micron scales (genes and cells) to anatomical structures at tissue and organ scales. Our approach addresses three key challenges:
- Comparing diverse modalities using the mathematical framework of image-varifold norms
- Managing massive data volumes through a multiscale strategy for computing non-rigid deformations efficiently
- Handling incomplete data via partial matching methods that map full-brain references to sparsely sampled sub-volumes
We introduce xIV-LDDMM, a cross-modality image-varifold Large Deformation Diffeomorphic Metric Mapping toolkit. This framework enables efficient representation and registration of peta-scale datasets, bridging spatial scales from nanometers to millimeters.
These methods are detailed in:
- Stouffer KM, Chen X, Zeng H, et al. xIV-LDDMM Toolkit: A Suite of Image-Varifold Based Technologies for Representing and Mapping 3D Imaging and Spatial-omics Data Simultaneously Across Scales. Submitted. 2025. PubMed
- Stouffer KM, Trouvé A, Younès L, et al. Cross-modality mapping using image varifolds to align tissue-scale atlases to molecular-scale measures with application to 2D brain sections. Nat Commun. 2024 Apr 25;15(1):3530. PubMed
