Alternative splicing drives transcriptomic diversity critical for cell identity and function, yet the field lacks tools to quantify the multidimensional nature of isoform complexity at single-cell resolution. This analytical gap limits our understanding of how isoform regulation governs differentiation, disease, and therapeutic responses.
We present ScIsoX, the first framework for systematic, multidimensional analysis of isoform-level transcriptomic complexity. ScIsoX introduces key innovations: (1) a novel Single-Cell Hierarchical Tensor (SCHT) for efficient data organisation; (2) seven core metrics that capture isoform complexity across biological scales; (3) statistical methods for gene complexity classification; and (4) an interactive visualisation ecosystem.
Applied to datasets from mouse hematopoiesis, brain development, and human peripheral blood, ScIsoX uncovers new biological insights, including coordinated isoform co-expression (e.g., Sox17, MS4A1), cell-type-specific isoform switching, and complexity dynamics distinguishing developmental from mature states.
In summary, by moving beyond differential transcript usage, ScIsoX opens new avenues for exploring transcriptomic complexity in health and disease.