This study introduces Fullscope-seq, which adds long-read sequencing of concatenated cDNA to Stereo-seq to place full-length isoforms at single-cell positions across large macaque brain sections. Barcode recovery improved (Fig. 1e), and thousands of genes showed differential isoform usage across layers, cell types and regions. Layer differences tracked cell composition, whereas many subregional differences were attributed to spatial context independent of cell type (Fig. 4g).
Alternative hypothesis: Part of the subregional spatial-context effect is technical rather than regulatory. Differences in RNA integrity, permeabilization or transcript spillover between neighbouring cells would shift isoform ratios for every cell type in an area at once, the pattern attributed to spatial context for RTN1 (Fig. 4i–k). Isoform-usage calls rely on read-level chi-squared tests within one animal, which give such shifts strong significance; in a second section from the same animal, the RTN1 difference between areas 1/2 and PEa is smaller (Extended Data Fig. 7g).
Ideally, subregional effects would be tested in sections from several animals, with RTN1-specific in situ hybridization in areas 1/2 and PEa. Or, at minimum, checking whether spatial-context genes are enriched for alternative start and end sites, the events most sensitive to RNA integrity, would show how much is technical.