Spatiotemporal analysis of human intestinal development at single-cell resolution

Abstract

Development of the human intestine is not well understood. Here, we link single-cell RNA sequencing and spatial transcriptomics to characterize intestinal morphogenesis through time. We identify 101 cell states including epithelial and mesenchymal progenitor populations and programs linked to key morphogenetic milestones. We describe principles of crypt-villus axis formation; neural, vascular, mesenchymal morphogenesis, and immune population of the developing gut. We identify the differentiation hierarchies of developing fibroblast and myofibroblast subtypes and describe diverse functions for these including as vascular niche cells. We pinpoint the origins of Peyer's patches and gut-associated lymphoid tissue (GALT) and describe location-specific immune programs. We use our resource to present an unbiased analysis of morphogen gradients that direct sequential waves of cellular differentiation and define cells and locations linked to rare developmental intestinal disorders. We compile a publicly available online resource, spatio-temporal analysis resource of fetal intestinal development (STAR-FINDer), to facilitate further work.

Keywords: congenital disease; gene expression; human development; human developmental cell atlas; intestinal crypt; intestinal development; mesenchymal cells; single-cell RNA-sequencing; spatial transcriptomics; stem cells.

Graphical abstract

Figure 1

Generation of a spatio-temporal transcriptional atlas of human intestinal development (A) Overview of study design for intestinal development atlas. (B) scRNA-seq experiment sample overview dot-plot depicting sample distribution across location, developmental time and high-quality post-QC cells recovered per sample. (C) UMAP embedding of single cell transcriptomes of cells from 9 different compartments. (D) Markers of tissue compartment specific genes used for cell annotation shown as fraction of expressing cells (circle size) and mean expression (color) of gene markers (columns) across compartment (rows). (E) UMAP embedding overlay showing the location distribution across all compartments. (F) UMAP embedding overlay showing the gestational age (PCW) distribution of single cells. (G) Partition-based graph abstraction (STAR Methods) of 101 cell clusters identified in scRNA-seq data (colored by compartment, line representing weight of interaction, legend for cell cluster annotation Table S1)

Figure S1

Overview of hashing methodology, sample cell-of-origin assignments and pool batch correction, related to Figure 1 and STAR methods (A) Hemotoxylin and Eosin (H&E) staining of intestinal sections demonstrating morphology of samples spanning the times and locations included in transcriptomic atlas (representative images of ≥3 samples at specified location and similar (+-1pcw) timepoints, each at 20x magnification scale bar=180 μm). (B) Example distribution of B2M mRNA expression in single EPCAM+ cells from an early gestation (8 PCW) sample and late gestation (19 PCW) sample from the same pool (identical sequencing depth and sample preparation conditions) showing reduced B2M mRNA levels in early gestation. (C) Density plot showing the distribution of per cell gene detection rate across different cell compartments. Cells are further broken down into G2M&S Phase cells (dashed line) and G1-phase cells (solid line) based on cluster analysis, as cycling cells in the G2M/S-phase tend to have substantially larger total mRNA content. (D) t-distributed stochastic neighborhood embedding (tSNE) of cells from a representative EPCAM- pool based on their recovered hashing antibody profiles, colored by classification into singlets, doublets or unstained/negative cells following dehashing (STAR methods). (E) tSNE embeddings of EPCAM+ cells from a representative pool, showing embeddings based on TotalSeq antibody tags only (i), in-house labelled antibody tags (ii) and both tags (iii). Cells are colored by sample identities assigned from dual-tag labels. Arrows indicate relevant regions highlighting multiplets (top left, (i) panel); inability to discriminate cells from low gestation samples (Sample 1 and Sample 2) when using TotalSeq tags only (center arrow, (i) panel), while the in-house tag separates cells from Sample 1 (left arrow, (ii) panel) from Sample 2 (right arrow, (ii) panel) while some untagged/negative cells still remain (center arrow, (ii) panel). (F) tSNE overlay comparing TotalSeq (i) and in-house (ii) tag signal in late gestation samples, and early gestation samples (iii-iv) in double-tagged EPCAM+ pool cells. In-house tag signal is stronger than TotalSeq tag for early gestation cells, while being comparable for late gestation cells. (G) tSNE embedding of TotalSeq tag signal in late (i) and early (ii) gestation samples in EPCAM- pools shows similar tag recoveries, in contrast to EPCAM+ pool cells in (F). (H) Expression of compartment markers for stromal cells - THY1/CD90 (i) and epithelial cells - EPCAM (ii) shown as a tSNE overlay over embedding shown in E (iii). Arrows highlight regions of interest, where most unstained/unassigned cells in EPCAM+ pools are either EPCAM- due to poor cell quality or are non-epithelial contaminants expressing stromal or immune markers (center arrow, panels i-ii). Non-epithelial contaminant cell sample-of-origin can be resolved in many cases in double-tagged pools, for instance stromal cells in central region of the embedding in other samples (Figure S1E iii). (I) Example of pool effect batch correction described in STAR methods. A UMAP embedding shows epithelial cells and their hashed pool prior to pool effect correction (i) and post-correction (ii).

Figure S2

Transcription factor regulatory networks and receptor-ligand interactions in scRNA-Seq of fetal intestinal development, related to Figure 1 (A) Hierarchical clustering of all cell types based on TF module scores shows a cell type TF “decision tree”. Branches are colored by compartment. Up to top two TFs most discriminative of a branch split are shown as labels (full data for each split is provided (Fawkner-Corbett et al., 2020). (B) UMAP overlay of selected TF module AUC scores in single cells across all compartments, demonstrating gene modules with compartment specific regulation of intestinal epithelium (ARID3A), fibroblasts (TCF21) and proximal and distal epithelial discriminating FOXD1. (C) Heatmap summarizing total cluster pairwise paracrine receptor ligand interactions. The frequency of interactions (row-wise and column wise cell type where color represents frequency; color bar indicates cluster compartment. Cluster numbers correspond to graph-abstraction in Figure 1G and Table S1)

Figure 2

Spatio-temporal analysis of intestinal development with ST and scRNA-seq integration (A) UMAP plot of spot transcriptome clusters from each slide shown on left; clusters are visualized on tissue covered slide areas (left center). Integration with scRNA-Seq cell type annotations are shown on the right, with tissue morphology of the region shown right center for 12 PCW TI (i), 12 PCW colon (ii), 19 PCW colon (iii) and adult colon (iv) slides. All H&E images are from selected areas of ST slides from the following tissue sections: A6 (i), A8 (ii), A4 (iii) and A1 (iv). Full images are available in (Fawkner-Corbett et al., 2020). (B) Validation of ST method by comparison of adult intestinal tissue spots with histological landmarks and known related single genes – crypt top colonocyte transcriptomic signature near crypt tops (top left, left-center) and myofibroblast signature near muscularis mucosa (top right, right-center); expression of known immune cell marker PTPRC/CD45 in spots covering submucosal lymphoid follicle (bottom left, left-center) and expression of RET at myenteric plexus (bottom right, right-center). All H&E images are from selected areas of ST tissue section from section A1, H&E reference image is repeated for clarity (top). Full image is available in (Fawkner-Corbett et al., 2020). (C) (i) Pairwise cell type prediction signal correlation heatmap in adult ST spots. Non-significant correlations (<0.05 adjusted p value) are shown in white; color bar indicates Pearson’s r value. Red boxes highlight selected biologically relevant correlation groups. (ii) Heatmap showing distance-smoothed expression of significant distance-varying genes detected in adult ST slide. Vertical break indicates muscularis mucosa/distance score of zero and spots in the submucosa are assigned a negative distance score while spots in the mucosa are assigned a positive distance score. Two broad gene clusters are assigned by cutting hierarchical clustering tree, dividing the gene groups into mucosa and sub-mucosa specific expression groups. Selected GO BP terms enriched in each cluster are shown. (iii) Selected cell type prediction distribution over distance/depth score (inset and legend: ST slide overlay showing distance measures from muscularis mucosa used to assign each spot a distance gradient colored by depth score overlaid over H&E image from ST section A1, full image available in (Fawkner-Corbett et al., 2020) in adult ST showing sequential distribution of cell types, predicted using adult single cell references from (Parikh et al., 2019, GEO: GSE116222) and (Kinchen et al., 2018, GEO: GSE114374) and (Smillie et al., 2019, DUOS-000110) (D) As in (C)(iii), distance/depth score applied to fetal ST slides (inset and legend: ST slide overlay showing distance measures from serosa used to assign each spot a distance gradient colored by depth score), showing selected cell type distribution across tissue depth from serosal membrane to lumen in samples from 12 PCW colon (i) and 19 PCW colon (ii). Inset spot overlay is shown over selected areas from H&E images from ST sections A8 (i) and A4 (ii). Full images are available in (Fawkner-Corbett et al., 2020).

Figure 3

Cataloguing in utero epithelial maturation and crypt development (A) UMAP plot visualizing epithelial compartment populations (i) and epithelial cell distribution based on location (ii) and developmental time course (iii). (B) Dot plot of epithelial cluster markers, with color indicating average expression within cluster and dot size indicating percentage of cells within cluster expression the gene. (C) Selected population abundance changes over developmental time course shown as bar plots. Wilcox rank test, p-value < 0.05 ^∗; p-value < 0.01 ^∗∗; p-value < 0.001^∗∗∗; n.s = not significant. For location-specific clusters, only location-matched samples were considered. Error bars represent standard error of the mean (SEM). (D) Violin plots showing expression of selected time-course varying genes in distal and proximal stem cells. (E) UMAP overlay visualizing expression of GATA4 in epithelial cells (i) and representative images of SI sections from 10, 17 and 22PCW embryonic tissue stained for GATA4 by immunohistochemistry (IHC) (n = 3 for each individual image shown repeated on samples +-1pcw PCW, 10x/20x magnification scale bar=360/180 μm) (ii) (F) Representative images of SI sections from 10 and 17 PCW embryonic tissue stained for Transferrin (TF) by IHC (n = 3 on samples +-1pcw to example image, 20x/100x magnification scale bar=180/40μm) (G) Representative images of colonic sections from 10, 12 and 17 PCW embryonic tissue and adult colonic tissue stained for LGR5 expression by single molecule in-situ hybridization (sm-ISH) (n=3 on samples +-1pcw in fetal images, 20x/100x magnification scale bar=180/40μm) (H) UMAP overlay of ASCL2 TF module AUC score over developmental time course in epithelial cells. (I) Representative images of colonic sections from 10 and 15 PCW embryonic tissue stained for BEST4 by IHC (n=3 on samples +-1pcw to example image, 20x/40x magnification scale bar=180/90μm respectively).

Figure S3

Epithelial compartment fetal intestinal development, related to Figure 3 (A) Developmental trajectory analyses of epithelial compartment cells using Monocle algorithm (i) shown over UMAP embedding, colored by pseudotime and RNA velocity estimates (ii) shown over UMAP embedding, colored by cell clusters. (B) Volcano plot (i) showing differentially expressed genes between colonic and TI stem cells. Selected genes are labelled. (C) Interacting cell network plots showing cell type cross talk via specific receptor-ligand pairs. Ligand source clusters are indicated as circles, receptor target cells as squares; autocrine interactions are shown as diamonds. Edges color indicates interaction score, node color ligand or receptor cell type specificity and node size indicates percentage of cells expressing ligand or receptor in the cluster. (D) Volcano plot (left) showing differentially expressed genes between stem cells and stem-like progenitor cells. Selected genes are highlighted. Violin plots showing stem cell-specific LGR5 expression and progenitor specific VTN expression are shown on the right. (E) Goblet (left) and enteroendocrine (right) population abundance changes over developmental time course shown as bar plots. Wilcox rank test, p-value < 0.05 ^∗; p-value < 0.01 ^∗∗; p-value < 0.001^∗∗∗; n.s = not significant. Error bars represent standard error of the mean (SEM). (F) Dot plot (i) heatmap showing selected epithelial secretory cell sub cluster markers. Points are scaled by percentage of cells with at least minimal (>0) detection of marker within the cluster and colored by mean cluster expression. Secretory sub-clusters are visualized as a UMAP embedding (ii), with overlays of developmental time point (iii) and location (iv). (G) BEST4/OTOP2 population abundance remains the same over developmental time course shown as bar plots (i). Wilcox rank test, p-value < 0.05 ^∗; p-value < 0.01 ^∗∗; p-value < 0.001^∗∗∗; ns = not significant. (ii) Circos plot visualizing putative cross-talk between BEST4/OTOP2 cells and Inhibitory Motor Neurons. Error bars represent standard error of the mean (SEM).

Figure 4

Coordinated development of mesenchymal and endothelial compartment cell (A) UMAP visualization of endothelial compartment subclusters. (B) Dot plot heatmap showing selected pericyte cell sub-cluster markers. Points are scaled by percentage of cells with at least minimal (>0) detection of marker within the cluster and colored by mean cluster expression. (C) (i) Mature myofibroblast population abundance changes over developmental time course shown as bar plots. Wilcox rank test, p-value < 0.05 ^∗; p-value < 0.01 ^∗∗; p-value < 0.001^∗∗∗; ns = not significant. Error bars represent standard error of the mean (SEM). (ii) Partition-based graph abstraction showing expression of myofibroblast marker RSPO2. (D) ST spot overlay of cell type predictions for myofibroblast cells and S3 transitional at 19 PCW. All H&E images and corresponding spot overlays show selected area (image repeated for clarity) from ST tissue H&E image section A4. Full image is available in (Fawkner-Corbett et al., 2020). (E) ENS and glial progenitor population abundance changes over developmental time course shown as bar plots. Wilcox rank test, p-value < 0.05 ^∗; p-value < 0.01 ^∗∗; p-value < 0.001^∗∗∗; ns = not significant. Error bars represent standard error of the mean (SEM). (F) Dot plot heatmap showing selected muscle cell sub-cluster markers. Points are scaled by percentage of cells with at least minimal (>0) detection of marker within the cluster and colored by mean cluster expression. (G) (i) UMAP overlay visualizing fibroblast sub-populations. UMAP overlay showing the distribution of fibroblast cells over developmental time (ii) and location (iii). (H) Co-localization of S3 progenitor cells and venous (CP) cells in 12 PCW ST slide (i); Co-localization of S3 marker C7 with large vessels in adult ST slide (ii) and localization of S3 EBF+ (iii) and S3 HAND1+ (iv) cells to areas surrounding vessels in adult slides. H&E image spot overlays show selected areas from ST tissue H&E image sections A8 (i) and A1 (ii-iv). H&E reference image repeated for clarity at each zoomed in location. Full images are available in (Fawkner-Corbett et al., 2020).

Figure S4

Development of endothelial and pericyte compartment cells, related to Figure 4 (A) Selected endothelial population abundance changes over developmental time course shown as bar plots. Wilcox rank test, p-value < 0.05 ^∗; p-value < 0.01 ^∗∗; p-value < 0.001^∗∗∗; ns = not significant. Error bars represent standard error of the mean (SEM). (B) ST spot cell type predictions of endothelial cluster 2 from single cell reference in (Kinchen et al, 2018) (top) and large venous single cell signature from scRNA-seq data here shown in adult slides (bottom). Areas with vessels are zoomed in for clarity. All H&E images and corresponding ST spot overlays are plotted over ST H&E section A1, reference H&E images are repeated for clarity (top and bottom zoom). Full image is available in (Fawkner-Corbett et al., 2020). (C) Interacting cell network plot (i) showing cell type cross talk via specific receptor-ligand pair, ANGPT2 and TIE1. Ligand source clusters are indicated as circles, receptor target cells as squares; autocrine interactions are shown as diamonds. Edges color indicates interaction score, node color ligand or receptor cell type specificity and node size indicates percentage of cells expressing ligand or receptor in the cluster. (ii) Representative images of late (19–22 pcw) colonic sections stained for ANGPT2 protein by IHC with positive expression around vessels observed in the lamina propria (LP) and serosa (n=4 experiments on individual samples 19–22pcw, 20x/100x magnification, scale bar=180/40μm). (D) Time course abundance changes in S- and G2M- phase pericytes, highlighting proliferating dynamics of pericyte compartments. Abundance changes in cycling and non-cycling counterparts of pericyte progenitors, WNT6+ pericytes and contractile pericytes suggest these represent earlier proliferative cell states, while other cell types represent more differentiated cell phenotypes. Error bars represent standard error of the mean (SEM). (E) Trajectory analysis using Monocle algorithm of fibroblast, pericyte, muscularis and myofibroblast compartment cells highlights differentiation of pericytes and myofibroblasts from S1 (marked by ADAMDEC1) and S3 (marked by OGN) like fibroblast progenitors respectively.

Figure S5

Development of muscularis and neural compartments, related to Figure 4 (A) (i) UMAP visualization of neural compartment subclusters. (ii) ST spot overlay of cell type predictions of neuroendocrine (1) cells in 19 PCW ST slide. All H&E images and corresponding ST spot overlays in (ii) are plotted over selected regions of ST H&E section A4. Full image is available in (Fawkner-Corbett et al., 2020). (B) UMAP visualization of muscularis compartment cell subclusters. (C) Selected muscle population abundance changes over developmental time course shown as bar plots. Wilcox rank test, p-value < 0.05 ^∗; p-value < 0.01 ^∗∗; p-value < 0.001^∗∗∗; n.s = not significant. Error bars represent standard error of the mean (SEM). (D) ST spot overlay of MYH11 expression (top) and MM and OM cell type predictions (bottom) localizing to distinct layers in 19 PCW colon ST slide. All H&E images and corresponding ST spot overlays in are plotted over selected regions of ST H&E section A4. Full image is available in (Fawkner-Corbett et al., 2020). (E) UMAP overlay showing muscularis-specific KLF7 and TWIST2 TF module AUC score distribution.

Figure S6

Development of fibroblast compartment, related to Figure 4 (A) Dot plot heatmap showing selected fibroblast cell sub cluster markers. Points are scaled by percentage of cells with at least minimal (>0) detection of marker within the cluster and colored by mean cluster expression. (B) Developmental trajectory analyses of fibroblast compartment cells using Monocle algorithm shown over UMAP embedding, colored by pseudotime (left). UMAP overlay (right) of NR2F1 TF module AUC scores, delineating S1/S2 type cells from S3 type fibroblasts. (C) Time course changes in the proliferating S- and G2M- phase cells in the fibroblast compartment. S3 progenitor population constitutes the most abundant and the most enriched (data not shown) over G1-phase cells population. Error bars represent standard error of the mean (SEM). (D) Circos plot visualizing putative cross-talk between S3+ HAND1+ cells and Arterial and Venous endothelial cells. (E) ST adult slide spot overlay of expression of receptor-ligand pair CEACAM1 and CEACAM5, showing significant co-localization of these molecules (i). ST slide overlay of LRP1 and HSPG2 receptor-ligand pair in adult ST slide (ii) and 12 PCW ST slide (iii). All H&E images and corresponding ST spot overlays in are plotted over selected regions of ST H&E sections A1 (i-ii) and A6 (iii), reference H&E images repeated for clarity. Full images are available in (Fawkner-Corbett et al., 2020).

Figure 5

Intestinal morphogen gradients in specific cell types and spatial locations (A) Graph visualization of morphogen molecule STRING interactome. Communities enriched for EGFR, FGF, Hedgehog, HIPPO, NOTCH, RTK, TGF-beta and WNT signaling pathways are highlighted in dashed ellipses. Nodes are colored by scRNA-Seq module. Nodes unassigned to a module are shown in grey (NA). (B) Individual morphogen module overview shown as a module score overlay in ST spots in slides at 19 and 12 PCW colon and a dotplot showing module gene expression at compartment level for ST module 3 which captures endothelial, pericyte and fibroblast morphogens (i) and scRNA-Seq module 8, which captures epithelial morphogens (ii). All spot overlays shown are plotted over ST H&E sections A4 (top) and A3 (bottom, rotated). Full images are available in (Fawkner-Corbett et al., 2020). (C) (i) Violin plot showing expression of RSPO3 over developmental time course in scRNA-Seq of muscularis compartment cells. (ii) Dotplot showing the expression of mesothelium and muscularis-specific morphogen scRNA-Seq module 5 genes (D) Individual morphogen module overview shown as module score overlay in ST spots in slides at 19 and 12 PCW colon and a doplot showing expression of fibroblast-specific morphogen spatial module 6. All spot overlays are plotted over ST H&E sections A4 (top) and A3 (bottom, rotated). Full images are available in (Fawkner-Corbett et al., 2020). (E) Violin plots depicting selected S2 genes (POSTN (i) and BMP3 (ii)) that show locational and time-course differences in expression. (F) Representative images of colonic sections from 11, 15 and 20 PCW embryonic tissue stained for F3 protein by IHC. Brown arrows within the 11 PCW sample indicate positive F3 staining below newly developing invaginations/hillocks in epithelium (n = 3 for each individual image with +-1pcw samples, 10x/20x magnification scale bar=360/180 μm)

Figure S7

Cell-type specific and spatial morphogen gradients in the developing intestine, related to Figure 5 (A) Heatmap visualizing the correlation structure of morphogen module genes expressed in ST data (i) and single cell RNA-Seq data (ii). The overlap between detected modules is visualized in (iii). (B) Individual morphogen module overview shown as a module score overlay in ST spots in slides at 19 and 12 PCW colon and a dotplot showing module gene expression at compartment level for scRNA-Seq module 11, which captures largely pericyte and myofibroblast-derived morphogens. Spot overlays are plotted over ST H&E sections A4 (top) and A3 (bottom, rotated). Full images are available in (Fawkner-Corbett et al., 2020). (C) Interacting cell network plots showing putative cell type cross talk via specific receptor-ligand pairs, RSPO3/LGR5(i) and DLL1-NOTCH2(ii). Ligand source clusters are indicated as circles, receptor target cells as squares; autocrine interactions are shown as diamonds. Edges color indicates interaction score, node color ligand or receptor cell type specificity and node size indicates percentage of cells expressing ligand or receptor in the cluster. (D) Volcano plot visualizing differentially expressed genes between colonic and TI S2 populations.

Figure 6

Early intestine immune colonization and immune-stromal interactions (A) UMAP visualization of immune compartment cell clusters. (B) Dot plot showing the expression of key immune compartment cluster markers. (C) Barplots showing time course changes in selected immune compartment cluster abundance. Wilcox rank test, p-value < 0.05 ^∗; p-value < 0.01 ^∗∗; p-value < 0.001^∗∗∗; ns = not significant. Error bars represent standard error of the mean (SEM). (D) (i) Violin plot showing expression of GFRA3 in lymphoid associated glial (LA Glial) cells and other neural cells in cells from colon and TI. (ii) Barplot showing time course changes in LA Glial abundance in neural compartment in colon and TI. Wilcox rank test, p-value < 0.05 ^∗; p-value < 0.01 ^∗∗; p-value < 0.001^∗∗∗; ns = not significant. Error bars represent standard error of the mean (SEM). (E) Violin plots showing expression of CCL19 (i), CCL21 (ii) and CXCL13 (iii) in S4 fibroblast cells over developmental time course. (F) ST cell type predictions in spots overlaying a submucosal lymphoid aggregate in adult tissue slide showing scores for adult Stromal 4, T-cells, B-cells and Plasma cells in the bottom of slide A1 (i) and center of slide A2(ii). H&E images and spot overlays show selected regions from ST sections A1 (i) and A2 (ii). Full images are available in (Fawkner-Corbett et al., 2020). (G) Expression of CCR7 and CCL19 in ST adult slides. H&E images and all spot overlays shown are plotted over selected regions of ST H&E section A1, H&E image of zoomed in section repeated for clarity. Full image is available in (Fawkner-Corbett et al., 2020).

Figure S8

Immune colonization of developing intestine, related to Figure 6 (A) UMAP visualization of immune compartment cell cluster locational (i) and PCW (ii) distributions. (B) Circos plot visualizing putative cross-talk between ILC3s and S4 CXCL13+ cells. (C) Interacting cell network plots showing cell type cross talk via specific receptor-ligand pairs, IL7/IL7R, CCL21/CCR7 and CCL19/CCR7. Ligand source clusters are indicated as circles, receptor target cells as squares; autocrine interactions are shown as diamonds. Edges color indicates interaction score, node color ligand or receptor cell type specificity and node size indicates percentage of cells expressing ligand or receptor in the cluster. (D) S4 cell sub-population (left- S4 CCL21+; left- S4 CXCL13+) abundance changes over developmental time course in colon and TI samples shown as bar plots. Wilcox rank test, p-value < 0.05 ^∗; p-value < 0.01 ^∗∗; p-value < 0.001^∗∗∗; n.s = not significant. Error bars represent standard error of the mean (SEM). (E) Volcano plot (i) visualizing differentially expressed genes between S4 CXCL13+ and S4 CCL21+ cell subpopulations. Global expression of TNFSF11 (RANKL) is visualized as a graph abstraction overlay in (ii).

Figure 7

Application of in utero gene expression profiles to developmental disease (A) Table summarizing intestinal disease HPO phenotype terms and the number of associated genes which are at least minimally expressed in scRNA-Seq dataset or highly cell type specific. (B) Heatmap visualizing mean scaled cluster expression of cell-type specific disease genes summarized in (A). (C) Graph abstraction overlay of cluster expression of disease genes associated with (i) congenital diarrhea (SPINT2 & DGAT1) and (ii) Hirschsprung’s Disease (RET & SOX10). (D) Volcano plots highlighting top time-course varying disease-associated genes in muscularis and neural compartments. (E) Violin plots showing individual time-course varying disease-associated genes in (i) intestinal malrotation, showing HMGA2 expression over time in glial vs neuron cells (left) and HMGA2 expression over time in muscularis compartment in TI and colon (right); and (ii) omphalocele showing NXN expression over time in neurons.