A0994
Title: From single cells to spatial maps: Computational alignment of transcriptomic data
Authors: Yunlu Chen - Northwestern University (United States) [presenting]
Feng Ruan - Northwestern University (United States)
Ji-Ping Wang - Northwestern University (United States)
Abstract: Spatial transcriptomics (ST) measures mRNA transcripts at thousands of locations within tissue slices, revealing spatial variations in gene expression and cell types. However, ST data lacks single-cell resolution, necessitating computational methods to align single-cell RNA-seq (scRNA-seq) data with ST. A novel alignment method is developed with an accompanying Python package, NLSDeconv, based on non-negative least squares for efficient cell-type deconvolution and spatial mapping of ST data. Benchmarking against 18 existing deconvolution methods across various ST datasets demonstrates the approach's competitive statistical performance and superior computational efficiency. Building on this foundation, the method is extended to handle large-scale ST datasets, and the open-source Python package SLSDeconv is developed. Looking forward, temporal alignment approaches are explored for ST datasets collected across multiple developmental time points.
