Bioinformatics Seminar
Time: 11AM
Venue: Davis Auditorium and Online
27 May 2025
NAScaling single-cell chromatin analysis with regulatory element modules
Tim StuartGenome Institute of Singapore
Single-cell epigenomic assays offer new ways to understand how the activity of DNA regulatory elements may shape cellular states and fates. Recent advances in experimental methods now allow profiling of millions of cells, enabling a much more reliable quantification of these epigenomic states, particularly for rare cells. However, the analysis of such data is still impeded by two computational challenges. First, there is a lack of reusable features for epigenomic analysis, and each dataset requires de novo peak calling. This makes comparison of published datasets difficult as they cannot be directly compared. Second, current analysis methods cannot scale to process large single-cell epigenomic datasets, requiring excessively long runtimes and large memory resources. To address these challenges developed a novel approach for single-cell epigenomic analysis based on feature aggregation. We leveraged epigenomic data from thousands of published datasets to identify groups of co-accessible regulatory elements, which we term Regulatory Element Modules (REMO). We further developed an open-source software toolkit, implemented in Rust, that provides fast and memory-efficient quantification of single-cell epigenomic data. REMO enables accelerated analysis of single-cell epigenomic data with lower memory requirements, and provides reusable features applicable to a broad range of tissue types, offering a robust framework for consistent and scalable analysis of single-cell epigenomic data.