Mining Single-Cell and Spatial Transcriptomics Data for Alternative Splicing Events

Advances in single-cell and spatial transcriptomics are transforming the study of gene regulation and cellular function. Single-cell RNA sequencing enables researchers to profile the transcriptomes of individual cells, revealing heterogeneity that is obscured in bulk tissue analyses. Even cells that appear morphologically identical can exhibit distinct gene expression patterns, which can have important implications for development, disease, and immune responses. Spatial transcriptomics complements this approach by preserving the tissue architecture, allowing gene expression to be mapped back to specific cellular locations and local microenvironments. This provides critical context for understanding how cells interact within tissues and how spatial organization influences cellular behavior. Together, these technologies provide a powerful framework for linking cellular identity, function, and location. They are increasingly used to investigate developmental processes, cancer progression, immune responses, and neurological disorders, offering insights that were previously inaccessible using traditional genomic or histological approaches. Despite these advances, most computational methods developed for single-cell and spatial transcriptomics focus only on gene-level expression. While useful, this overlooks the complexity of alternative splicing - a key mechanism that allows a single gene to generate multiple transcript isoforms and protein variants. Alternative splicing dramatically expands transcriptomic diversity and has well-documented roles in cell differentiation, tissue organisation, and disease. Misregulation of splicing has been implicated in cancer, immune dysfunction, and neurological disorders, but its contribution at single-cell and spatial resolution remains poorly explored. Detecting and characterising splicing events in these new data types presents significant computational challenges. Single-cell and spatial transcriptomics datasets are massive, noisy, and sparse, making current splicing analysis tools inadequate. Existing algorithms are typically designed for bulk RNA-seq, and do not scale efficiently or perform well under the data characteristics of single-cell and spatial assays. There is therefore an urgent need for new computational approaches capable of efficiently handling large-scale datasets, resolving isoform variation, and producing biologically meaningful results. This research project will develop novel computational strategies for alternative splicing discovery in single-cell and spatial transcriptomics. Building on the Subread, featureCounts, and cellCounts frameworks established in our lab, the student will design new algorithms and software to detect, quantify, and visualise splicing patterns at cellular and spatial resolution. The focus will be on creating highly efficient, scalable, and robust computational methods that can process large datasets and integrate seamlessly into modern bioinformatics workflows. The resulting tools will enable new discoveries about the role of splicing in cellular identity, tissue organisation, and disease. The student will receive training in algorithm design, high-performance computing, and large-scale transcriptomics data analysis, working at the cutting edge of computational genomics. This project offers opportunities to collaborate with experimental researchers generating single-cell and spatial data, ensuring that the methods developed are widely applicable and impactful. Students interested in this project should have a strong computational background (e.g. software engineering, computer science, bioinformatics, statistics, or related fields). Prior experience with biological data is not required, but an interest in applying computational innovation to genomics will be important.