Fig. 3: Clustering concatenated scRNA-seq binarized data and scATAC-seq 10 K PBMC data.
From: Facilitate integrated analysis of single cell multiomic data by binarizing gene expression values
a UMAP visualization of the pre-binarized scRNA-seq data, colored by Leiden clusters with our re-annotated cell types. b UMAP of the scATAC-seq data, colored by re-annotated cell types. c Integrated clustering of the binary scRNA-seq data and the scATAC-seq data, colored by clusters. The number of BC clusters were chosen to match that from MOFA integration (Fig. 4). d Bubble plot showing the expression of the cell type markers previously used to annotate cell types. Note that the expression in all bubble plots was from pre-binarized data. e Alluvia plot showing the cell relationship among RNA cell types in (a) (left), the BC clusters in (c) (middle), and the ATAC clusters in (b) (right). f Cluster accuracy, as determined by the %s of cells belong to the dominant cell type determined in (a) or (b) for each of the integrated BC clusters (x-axis). Cell type annotations were derived using canonical markers and served as a standard for benchmarking method comparison rather than definitive biological truth. Source data are provided as a Source Data file.
