Fig. 6

MAESTRO outperforms the existing software in integrating scRNA-seq and scATAC-seq dataset. a Comparison of the cell-type label prediction score distribution after integration of the scRNA-seq and scATAC-seq using gene activity scores from MAESTRO, Seurat, snapATAC, and cicero. The comparisons were made on three independent datasets: PBMC from different donors, PBMC from the same donor, and BMMC from the same donor. Statistical significance was evaluated using Wilcoxon rank-sum test. b Comparison of the consistency between gene expression and gene activity scores from MAESTRO, Seurat, snapATAC, and cicero. The comparisons were made on three independent datasets. The x-axis represents the different cell types, and the y-axis represents the Spearman’s correlation coefficient between the expression level from scRNA-seq and gene activity score from scATAC-seq within each cluster. Only the top 2000 highly variable genes were used in the analysis