Fig. 5
From: iMAP: integration of multiple single-cell datasets by adversarial paired transfer networks
Applications of iMAP on CRC tumor-infiltrating immune cells. a Visualizations of major cell types and integration performance of iMAP. b Comparisons of the number of detected genes, transcription factors (TFs), ligand-receptor genes, and the percent of ribosome genes, noncoding RNAs between original 10x data, corrected 10x data, and Smart-seq2 data. c Comparisons of the dropout-ratio distribution between original 10x data, corrected 10x data, and Smart-seq2 data. d Re-analyzed of ILC subsets after batch effect removal. We showed new defined clusters and platform distribution over clusters. e New refined CD8+ T cell cluster with expression of key markers compared between original and corrected 10x data. f Significant ligand-receptor pairs (p < 0.05) defined between cDCs and T cells. g Significant ligand-receptor pairs detected between cDCs and Th1-like cells (CD4-CXCL13) by Smart-seq2 and corrected 10x data. h Selected specific pairs identified between cDCs and Th1-like cells by corrected 10x data. See also Additional file 1: Fig. S13