A: Confocal microscopy images of whole-mount mouse embryos at E9.5, E10.5, and E11.5 (left to right). PGCs are visualized via the Pou5f1-ΔPE-eGFP transgene (MGI:3057158) stained with anti-GFP antibody (ab13970). P indicates posterior of migratory stream; A indicates anterior. White arrows indicate the general direction of PGC migration by timepoint. B: UMAP plot of all cells profiled in this dataset colored by timepoint arranged using STITCH by timepoint (see methods - Batch correction). C: UMAP plots of all cells colored by log-normalized expression of marker genes relevant in PGC development. D: UMAP plot of all cells profiled in this dataset colored by annotated celltype. E: Matrixplot of genes with anteroposterior domain-specific expression at E9.5 and E10.5.

A: UMAP plot of PGCs downsampled to 1268 cells/timepoint colored by position-time of sample. B: UMAP plot colored by Pou5f1 gene expression. C: UMAP plot colored by Dazl gene expression. D All initial and terminal states predicted from CellRank transition matrix analysis based on diffusion pseudotime plotted on UMAP representation of downsampled PGCs across timepoints. E: Matrixplot of fate probability of reaching identified terminal states by position-time identity of PGCs. F: Putative driver genes with expression dynamics correlated with terminal state 1. Color bar at top of plot indicates position-time of cells most highly expressing driver genes below; see color key in 2A. G: UMAP plot of human PGCs and somatic cells from Li et al. 2017 colored by gestational week. H: UMAP plot of human PGCs and somatic cells colored by annotated celltypes. I: Violin plots of log-normalized expression of selected genes with conserved expression patterns between mouse and human.

A: Comparison of intercellular communication networks enriched in E9.5 anterior cells vs. enriched in E9.5 posterior cells by normalized information flow computed with CellChat. B: Comparison of intercellular communication networks enriched in E10.5 anterior cells vs. enriched in E10.5 posterior cells computed with CellChat. Relative scaled amount of information flow, left; Normalized amount of information flow, right. C-H: Cell populations in grey text do not participate significantly in plotted signaling pathway based on CellChat analysis. C: ncWNT signaling pathway network within mouse E9.5 posterior. D: ncWNT signaling pathway network within mouse E10.5 posterior. E: ncWNT signaling pathway network within mouse E10.5 anterior. F: Human ncWNT signaling pathway network. G: Human Ephrin A signaling pathway network. H: Human Ephrin B signaling pathway network.

A: UMAP plot of E9.5 PGCs colored by anteroposterior position. B: Matrixplot of expression of selected differentially expressed genes between anterior and posterior E9.5 PGCs. C: All initial and terminal states predicted from CellRank transition matrix analysis computed with diffusion pseudotime plotted on UMAP representation of E9.5 PGCs. D: Matrixplot of fate probability of reaching identified terminal states by position identity of PGCs at E9.5. E: Gene expression trends between initial state 3 and the four identified terminal states for selected genes.

A: UMAP plot of E10.5 PGCs colored by anteroposterior position. B: Matrixplot of expression of selected differentially expressed genes between anterior and posterior E10.5 PGCs. C: All initial and terminal states predicted from CellRank transition matrix analysis computed with diffusion pseudotime plotted on UMAP representation of E10.5 PGCs. D. Matrixplot of fate probability of reaching identified terminal states by position identity of PGCs at E10.5. E: Gene expression trends between initial state 3 and the four identified terminal states for selected genes. Diamonds correspond to functional relevance of selected genes. F: Representative image of whole-mount immunofluorescence staining of Oct4GFP, and LEFTY1/2. Fluorescence from Oct4-ΔPE-eGFP reporter signal is amplified by anti-GFP antibody (ab13970) stain. G: Quantification of LEFTY1/2 signal intensity of PGCs in n=3 embryos, one of which is show in F. H: Representative image of whole-mount immunofluorescence staining of E-Cadherin, Oct4GFP, and pSMAD2/3. Fluorescence from Oct4-ΔPE-eGFP reporter signal is amplified by anti-GFP antibody (ab13970) stain.

A-B: Barplot of enriched terms from gene ontology analysis of differentially expressed genes between E9.5 and E10.5 PGCs. C-D: Barplot of enriched terms from gene ontology analysis of differentially expressed genes between E9.5 and E11.5 PGCs. E: Visualization of random walks through CellRank transition matrix computed with diffusion pseudotime of 100 cells randomly selected from the E9.5 posterior sample stepping through 100 time steps. Black dots indicate starting positions, yellow dots indicate final positions, and lines between dots connect the positions of subsequent steps along the random walk, colored from violet (early) to orange (late). E-F: Visualization of random walks through CellRank transition matrix computed with experimental time of 100 cells randomly selected from the E9.5 timepoint stepping through 100 time steps. Black dots indicate starting positions, yellow dots indicate final positions, and lines between dots connect the positions of subsequent steps along the random walk, colored from violet (early) to orange (late). G. Violin plots of log-normalized expression levels of key genes of interest identified from mouse migratory PGC dataset in PGC, gonocyte, gonadal soma, and epithelial cell populations from Li et al. 2017 human dataset.

A-D: Incoming signaling networks between somatic populations (x axis) and PGCs across developmental time and position. E) FGF signaling during E9.5 PGC migration is spatially distinct, with anterior cells receiving more FGF signal from the neural tube. F) NCAM signaling during E9.5 PGC migration is spatially distinct, with anterior cells receiving signal from most somatic niches. G) EPHA signaling during E9.5 PGC migration is spatially distinct, with anterior cells receiving signal from most somatic niches as well as cell intrinsic signal. H) EPHB signaling during E9.5 PGC migration is not spatially distinct, with anterior cells and posterior cells receiving signal from most somatic niches. I) EPHA signaling during E10.5 PGC migration is spatially distinct, with posterior cells exchanging signals most strongly with the coelomic epithelium. J) EPHB signaling during E10.5 PGC migration is spatially distinct, with posterior cells receiving signal from endothelial cells and neural progenitors.

A: E9.5 mouse PGC samples before and after cell cycle phase regression labeled by predicted cell cycle phase. B: Barplot of enriched terms from gene ontology analysis of differentially expressed genes between E9.5 anterior and posterior migrating PGCs. C: Putative driver genes with expression dynamics correlated with terminal state 1.

A: E10.5 mouse PGC samples before and after cell cycle phase regression labeled by predicted cell cycle phase. B: Nodal expression is strongest in E10.5 PGCs. Top, UMAP plot showing Nodal gene expression among E10.5 PGCs and not from other cells. Bottom, violin plot of log-normalized Nodal expression in PGCs at each surveyed timepoint. C: Barplot of enriched terms from gene ontology analysis of differentially expressed genes between E10.5 anterior and posterior migrating PGCs. D. Expression of log-normalized NODAL mRNA across germ cells by timepoint in Li et al. 2017 human dataset. E: Putative driver genes with expression dynamics correlated with terminal state 2. F) Correlation plot of Lefty1/2 and Nodal genes with Y-chromosome genes, showing little correlation between Lefty/Nodal expression and Y-chromosome genes.