Fig. 5: Transcriptomic landscape of early L. major lesions.

FFPE sections from LC001, LC003 and LC008 were processed for Visium spatial transcriptomics with cell deconvolution performed using cell2location based on skin cell types identified by Reynolds et al.³⁴. a,b, Clustering of spots reveals 13 clusters in UMAP space (a) and with discrete spatial locations (b). Cluster locations are mapped to lesion and adjacent sections from LC001. Mapping to LC003 and LC008 and additional sections from LC001 are shown in Extended Data Fig. 4. c, Proportion of spots attributed to each cluster in healthy and lesion tissue. Data are pooled across all participants/sections. d, Heatmap representation of cellular abundances by cluster as determined by cell2location using the Reynolds et al.³⁴ reference dataset. Scale represents predicted 5% quantile abundances (q05 = 5% quantile values of the posterior distribution). e, Box and whisker plots representing cellular abundances/spot as in d for lesion core (cluster 2) and ulcer (cluster 7). Data are shown for the top 20 most abundant cell types. n = 20,241 spots derived from four sections from each of three participants. Box bounds show interquartile range (IQR) from the 25th to the 75th percentile; whiskers show the smallest and largest values within 1.5× the IQR from the lower and upper quartiles; and outliers are shown as data points outside the whiskers. f, Pairwise Pearson’s correlations are represented as a correlation plot between cell types to infer spatial co-localization. g, Volcano plot of differentially expressed genes (log₂FC > 1.5 and FDR = 0.05) comparing lesion core with ulcer. Statistical analysis was performed using two-sided Wilcoxon rank-sum test with Bonferroni correction for multiple comparisons.