Speaker: Fangming Xie, Ph.D.
IDRE Fellow
Institute for Quantitative & Computational Biosciences
University of California Los Angeles

Location: Virtual
Recording of the presentation

Abstract: Biology is undergoing a revolution of information explosion. In genomics, a single experiment measures the expression levels of thousands of genes in hundreds of thousands of cells. Meanwhile, scalable computational tools were used to distill understanding from complex data. The combination of high-throughput experiments and analyses has transformed research paradigms and advanced our understanding in many areas, including cell type taxonomy, developmental trajectories, and disease biomarkers. Yet, we are still far from gaining the throughput to measure and analyze transcriptomes (cells’ full gene expression profiles) at whole-organ scale in organisms that are much smaller than humans. Existing methods, including single-cell RNA-seq and spatial transcriptomics, will take years to measure all cells (n=~100 million) in a mouse brain, before any computational analyses to reveal cell types and their spatial organizations.

Rather than performing experiments first and analyzing data second, we propose an integrated approach where computational dimensionality reduction methods are used to design smarter experiments that maximize the information content of spatial transcriptomics experiments. Using a variety of methods, from principal component analysis, non-negative matrix factorization, to neural network, we extracted from transcriptome data (num. dimension > 10,000) low-dimensional (n=24) latent features, with each latent feature being a linear combination of many genes. We designed spatial transcriptomic experiments to directly measure those latent features in situ in the mouse brain. Preliminary data from one full coronal section of the mouse brain suggests the empirically measured latent features reveal rich spatial organization that matches known brain anatomy and cell types. To demonstrate its throughput, we are working on realizing its potential of enabling whole-organ scale spatial transcriptomics.

About the speaker: Dr. Xie is a postdoc in the Department of Chemistry and Biochemistry at UCLA. Prior to joining UCLA, Fangming received a BS degree in Physics at the University of Science and Technology of China. He visited UCLA as an undergraduate student for a summer, working with Dr. Robijn Bruinsma and Dr. William Klug to model the self-assembly of viral capsids. He then pursued a PhD at UCSD where his research focused on integrative analyses of single-cell transcriptomes and epigenomes of brain cells. Fangming loves neuroscience, genomics, and physics. He believes many parts of these disciplines can be brought together to sharpen our tools and advance our understanding of the brain.