Fig. 3. Inference of cell-type specific networks of mouse cellular reprogramming data.

a UMAP of LIGER cell clusters on the scATAC-seq data and scRNA-seq data. b UMAP depicting the sample labels of the scATAC-seq and scRNA-seq data from mouse cellular reprogramming. c The distribution of samples in each LIGER cluster. d The distribution of LIGER clusters in each sample. e Inferred lineage structure for scMTNI linking the 7 cell clusters with scRNA-seq measurements. f F-score of top 1k edges in predicted networks of scMTNI, scMTNI+Prior, INDEP, INDEP+Prior, SCENIC and CellOracle compared to three gold standard datasets: ChIP, Perturb and Perturb+ChIP. The top boxplots show the F-scores of n = 7 cell clusters, while the bottom heatmaps show FDR corrected t-test comparing the F-scores of the row algorithm to that of the column algorithm. The two-sided paired t-test is conducted on F-scores of n = 7 cell clusters for every pair of algorithms. A FDR < 0.05 was considered significantly better. The sign < or > specifies whether the row algorithm’s F-scores were worse or better than the column algorithm’s F-scores. The color scale is specified by − log(FDR), with the red color proportional to significance. Non-significance is colored in gray. In the boxplot, the horizontal middle line of each plot is the median. The bounds of the box are 0.25 quantile (Q₁) and 0.75 quantile (Q₃). The upper whisker is the minimum of the maximum value and Q₃ + 1.5*IQR, where IQR = Q₃ − Q₁. The lower whisker is the maximum of the minimum value and Q₁ − 1.5*IQR. g Pairwise similarity of networks from each cell cluster using F-score on the top 4k edges. Rows and columns are ordered based on the dendrogram created using the F-score similarity. Source data are provided as a Source Data file.