Extended Data Fig. 2: IRISeq quality control and performance comparison.

(a) The top panel displays the distribution of unique bead-bead interactions per receiver bead. The bottom panel shows the distribution of connected sender beads per receiver bead, with the median indicated by a dashed line in both panels. (b) UMAP plots compare the spatial distribution of receiver beads from a small-scale IRISeq experiment (left) with 10x Visium data (right), colored by the expression of region-specific markers Ttr (ventricle region, top) and Hpca (hippocampus region, bottom). (c) UMAP plot showing the integrated gene expression profiles from IRISeq receiver beads and 10x Visium spots, colored by assay type. (d) UMAP plot illustrating spatial distribution colored by annotated brain regions, similar to Fig. 1g. (e) UMAP plot showing the reconstructed spatial distribution of receiver beads from IRISeq, with beads colored according to the number of unique interaction oligos with a selected sender bead, validating the spatial reconstruction pipeline. (f) Histograms depicting Euclidean distances (in Supplementary Fig. 3e) between pairs of receiver beads connected to the same sender bead (top) versus distances between randomly selected receiver beads (bottom). (g) Running UMAP across multiple seeds. UMAP reproducibility across different seeds. Violin plots show reproducibility metrics across six UMAP runs. Procrustes RMSE (blue) measures global alignment error after orthogonal Procrustes adjustment, while kNN overlap (green = k15, red = k30) reflects local neighborhood stability. (h) Depiction of a large-area IRISeq experiment using a 1.5 cm x 1.5 cm bead array to profile two entire brain sections. (i) UMAP plot of receiver beads from the large-area IRISeq experiment, colored by the number of unique transcripts detected per bead. (j-k) UMAP plots displaying gene expression clusters (j) and reconstructed spatial distributions (k) of receiver beads from the large-area experiment, colored by annotated brain regions. (l) Quality-control metrics showing the distribution of total UMI counts (left) and detected gene counts (right) per 30 µm bin following filtering to retain bins with ≥700 UMI counts. (m) UMAP visualization of kidney gene expression transcriptomes colored by regional type. (n) Spatial reconstruction of the same dataset showing the spatially resolved distribution of identified regions. (p-r) Spatial plots of gene expression identified to be regionally specific. (s) Dot plot showing regional specific gene expression markers. Schematic in h created in BioRender; Cao, J. https://biorender.com/i2ew4vq (2026).