Fig. 5. Comparison of lineage projection in vitro with that of in vivo.

a UHC of the averaged transcriptomes of in vivo (M, T1) and in vitro cell clusters (hiPSCs, iMeLCs, hPGCLCs, MLCs, TCs, T1LCs) defined in Figs. 1a, 4a. b PCA for prospermatogonial stage GCs (M, MLCs, TCs, T1, and T1LCs). Color codes for cell clusters are indicated. c Number of DEGs and coefficient of determination (r²) for pairwise comparisons between in vivo (M [left] or T1 [right]) and in vitro clusters (hPGCLCs, MLCs, TCs, and T1LCs). DEGs are defined as genes with more than fourfold differences between two groups (mean log2[TPM+1] > 2, FDR < 0.01). r² are calculated using all annotated genes (18447 genes). d Scatter plot comparison of the averaged values of gene expression between T1 and T1LCs. Blue, genes higher in T1; red, genes higher in T1LCs (more than 4-fold differences [flanking diagonal lines], mean log2[TPM+1] > 2, FDR < 0.01). Key genes are annotated and the number of DEGs are indicated. Representative genes and their GO enrichments for genes higher in T1LCs are shown on the right. e Heatmaps of the expression of markers for “migrating” (172 genes), “mitotic” (306 genes), and “mitotic-arrest” (1037 genes) human male FGCs (defined by Li et al.¹¹) in the respective cell clusters defined in this study (left) and the indicated male FGC types defined by Li et al.¹¹ (right). Using these markers, hierarchical clustering was performed for cell clusters defined in this study (left). GO terms for the respective markers are shown on the right. f Heatmap of the averaged expression values of 316 genes in indicated cell types defined by Yamashiro et al.¹⁵ (top). Among markers for T1LCs defined in Fig. 4e, 1177 genes were also annotated in the dataset by Yamashiro et al.¹⁵. Among them, 316 genes upregulated in ag120 AG^+/−VT⁺ oogonia-like cells relative to all other cell types (more than twofold difference) are shown (top). GO analysis for these 316 genes are also shown (bottom). g Heatmap of the averaged expression in the indicated cell types for 110 genes highly expressed in T1LCs but with weak or no expression in ag120 AG^+/−VT⁺ oogonia-like cells. To enable comparison between two different scRNA-seq platforms, RPM values of the data from Yamashiro et al. were adjusted using a polynomial regression curve (see “Methods”). Among 1,177 DEGs for T1LCs (Fig. 4e), 110 genes showing high levels of expression in T1LCs (mean log₂[TPM+1] > 4) and low levels of expression in oogonia-like cells (adjusted log₂[RPM+1] < 2) are shown. GO analysis for these 110 genes are also shown (bottom). h Violin plot showing the expression levels of genes at the male-specific regions of the Y chromosome (MSY) in indicated cell clusters. These genes were identified the by multi-group DEG analysis in Fig. 4e (without cut-off by fold-change > 2). See also Supplementary Figs. 4, 6, 7, Supplementary Dataset 3–5.