IsoPrism predicts isoform usage from short-read transcriptomes through structure-guided competitive allocation.

IsoPrism predicts within-gene isoform proportions from short-read single-cell and spatial transcriptomes, achieving the highest mean dominant isoform accuracy and compositional similarity across five transfer scenarios.

• Why it matters: Short-read transcriptomics captures gene-level expression but often lacks resolution for isoform usage, limiting isoform-aware analysis in samples without matched long-read data.
• What they did: IsoPrism uses structure-guided competitive allocation with shared iterative updates to keep non-negative, gene-normalized proportions, supervised by individual long-read counts and group-level isoform usage.
• The result: In five transfer scenarios from three paired transcriptomic resources, IsoPrism produced the best dominant isoform accuracy, compositional similarity, and cluster-level allocation correlation among eight methods, and in a developing Mus musculus cortex atlas it revealed stage- and cell-type-associated isoform-usage patterns.

AI summary of the abstract · may contain errors · not medical advice · read the paper.