Seminar by Dr. Xiaoyu Song from Biostatistics at Centre for Quantitative Medicine, Duke-NUS Medical School

Recent advances in spatially resolved transcriptomics enable fine-grained identification of cell-cell interactions in spatially coordinated cells. We present QuadST, a highly robust and powerful approach for identifying cell-cell interaction changed genes (ICGs) based on single-cell spatially resolved transcriptomics data. QuadST constructs anchor-neighbor cell type integrated data matrix, and models the association between gene expression and the cell-cell distance dynamically at different distance quantile levels, and contrasts association signals across varying distances to identify ICGs. QuadST is highly robust as it maintains well-controlled type I error under the existence of unadjusted confounding factors, biased measured cell-cell distance, and gene-gene correlations. QuadST also achieves greater power than existing methods by modeling the quantitative impact of cell-cell distances on interactions. An application of QuadST to the Xenium profiled breast cancer data identified critical genes (CCL5 and LUM) involved in cancer and CD8+ T cell-cell interactions.

Dr. Xiaoyu Song is an Associate Professor of Biostatistics at Centre for Quantitative Medicine at Duke-NUS Medical School in Singapore. She graduated from Columbia University in 2015, and worked as Assistant/Associate Professor of Biostatistics at Mount Sinai between 2017-2023. Dr. Song’s primary research area is omics data analysis, focusing on developing novel statistical methods for association, integration, prediction, and network analysis of multiple -omics data. She has applied them for understanding molecular and cellular basis of complex human diseases like cancer. Dr. Song has published on prestigious statistical journals like JASA and biological journals like Cell.