A WGS-guided sc multiomic analysis integrates transcriptome and epigenome profiles of subclones. (A) Experimental overview of the patient cohort and available sequencing data. N/A indicates sample not available or sequencing yield <200 cells after quality control (QC). (B) UMAP embedding showing (left) scATAC- and (right) scRNA-seq data after QC, based on latent sematic indexing. Individual malignant PCs are colored according to patient identity. The bar plot on the right shows the number of cells per patient after QC. (C) Representative copy number profiles for the 22 autosomal chromosomes from (top) WGS, (middle) scRNA-, and (bottom) scATAC-seq of patient RRMM15. The color bar on the left indicates the respective subclonal population and sampling time point before relapse treatment (T1) and at subsequent relapse (T2). (Top) Copy number profile highlighting copy number gains (red), loss of heterozygosity (gray), and copy number losses (blue). Black represents a diploid copy number status. (Middle) Heatmap showing modified gene expression values generated using inferCNV.³⁵ Gains (red) and losses (blue) are highlighted. The scales were limited to 0.9 and 1.1. (Bottom) Heatmap for z-scores from scATAC-seq, using the identical color code as shown for WGS and scRNA-seq. Z-scores were limited to 3 and –3. (D) Pearson correlation between the subclone frequency identified via scRNA- (x-axis) and scATAC-seq (y-axis) for all analyzed patients with both modalities. Each time point was plotted separately, with each subclone being colored based on the patient. Linear regression line is depicted in black, in which the gray area marks the 95% confidence interval. The deviation between the subclone frequency between scRNA-seq and scATAC-seq was from 0% to 23.9% (mean deviation = 5.08%).