Please join us online for a talk by Dr. , Assistant Professor of Computational Genomics at UConn Health. He will present 鈥淭owards Location-Resolved Statistical Inference in Spatial Transcriptomics鈥� on Wednesday, September 9, at 1:30 pm Central Time. For access to this seminar, contact the Department of Biostatistics.
Dr. Song鈥檚 research focuses on developing statistical and computational methods for analyzing high-throughput omics data, particularly single-cell and spatial omics. He received his PhD in Bioinformatics from UCLA, his MS in Computational Biology from Harvard TH Chan School of Public Health, and his BS in Biological Sciences from Fudan University.
Abstract
Comparative analysis of spatial transcriptomics requires accurate alignment of tissue sections to a common coordinate system. Existing methods can be limited by alignment accuracy and scalability, while residual mismatch after alignment may make nominally corresponding locations biologically non-comparable and generate spurious differential expression signals. Here we present spAlignDE, an integrated framework for structure-guided spatial alignment and mismatch-aware local differential expression analysis. spAlignDE represents spatial structures as continuous fields and aligns them through shooting-based diffeomorphic registration, enabling both cross-sample and cross-modal alignment without requiring shared molecular features. Across cross-sample benchmarks against 12 methods, spAlignDE achieved the highest overall agreement in gene expression patterns and anatomical annotations and scaled to 20 MERFISH mouse brain sections comprising 1.45 million cells. After alignment, spAlignDE estimates local expression contrasts on a shared spatial grid and adjusts their uncertainty according to mismatch risk inferred from putatively stable genes and local observation density, with optional adjustment for cell-type composition. Simulations showed improved false-discovery control without systematic loss of power. Applications to mouse brain aging localized age-associated expression changes and shifts in T-cell spatial distributions, while analysis of kidney injury identified spatially restricted expression differences between normal and injured sections.