Fig. 1

SlideCNA methodology schematic and benchmarking on in silico Slide-seq-like data generated from snRNA-seq data of two MBC biopsies. a SlideCNA methodology involves calculating bead distances by a combined spatial- and expression-based pseudo-distance, binning beads based on pseudo-distance, and determining malignant clones based on CNA profiles. SlideCNA heat map (amplification > 1, deletion < 1), spatial plot of bins colored by assigned cluster, and boxplot of number of reads per bin after filtering for beads with > 300 counts across all genes for the in silico non-malignant-separated dataset (b) and non-malignant-mixed dataset (c) with UMI counts downsampled to 10%. Comparison of average SlideCNA CNA scores per chromosome arm of each SlideCNA-defined cluster for in silico data with non-malignant separation (d) and mixing (f) with UMI counts downsampled to 10% to the InferCNV profiles from the original snRNA-seq data. Pairwise Spearman correlation of CNA profiles as in d and f for in silico data with non-malignant separation (e) and mixing (g) with UMI counts downsampled to 10%