Tissular chromatin-state cartography based on double-barcoded DNA arrays that capture unloaded PA-Tn5 transposase

(Downloading may take up to 30 seconds. If the slide opens in your browser, select File -> Save As to save it.)

Click on image to view larger version.

Figure 4.
Figure 4.

Inferring chromatin-state transitions across the mouse embryo by integrating consecutive spatial epigenomic landscapes. (A, left) Scan of a DNA array (TRITC filter) hosting a cryosection of a mouse embryo (E11.5) and immunostained with an antibody targeting the histone modification H3K27-acetylation (H3K27ac). Notice the presence of Cyanin-3 (Cy3)-labeled DNA probes delimiting the DNA array composed of 32 × 32 interstitially printed probes. (Middle) DAPI staining. (Right) Digital map displaying the normalized read counts captured per physical position (SpExel) across the mouse embryo section. (B) t-SNE analysis allowing to stratify SpExel in six different clusters. (Right) Projection of the stratified clusters within the digital map. (C) Tissue–cell type gene marker association analysis performed per clusters identified in B. (D) Local enrichment signatures associated with six gene promoters presenting H3K27ac overrepresented counts. (E) In situ hybridization (ISH) and gene expression data (Allen Mouse Brain Atlas) for the genes Sox10 and Lhx5, revealing their spatial signature, which is coherent with the H3K27ac digitized view displayed in D. (F, left to right) Histone modification immunostaining (TRITC filter: H3K4me3, H3K27me3), nuclei revealed by DAPI, digital map of the normalized read counts across the tissue section, and local read count enrichment associated to the gene Wnt11. (G) Strategy for interrogating changes in chromatin histone modification signatures across the embryo: (i) overlay all three histone modification maps, (ii) generate a pseudomap in which SpExel of each digitized map are allocated to common pseudocoordinates, and (iii) interrogate for contiguous enrichment patterns similarity within tissue section but also across all three aligned sections. (H) Pseudomap obtained from the overlay of all three histone modification digital maps. (I) Promoter enrichment patterns associated to Wnt11 gene (more than five contiguous SpExel) retrieved in either of the histone modification maps. (J) Spatial gene promoter's coenrichment analysis for Wnt11 in the H3K27ac or H3K27me3 digital map. SpExel colored in red correspond to the location of Wnt11, whereas others correspond to other gene promoters sharing a similar spatial pattern (Tanimoto similarity index). Notice that these two spatial gene promoter's coenrichment maps present distinct spatial localizations. (K) Heatmap displaying the co-occurring promoter enrichment patterns between all three histone modifications. Six chromatin co-occurring states were identified and functionally associated to either active, repressed, or bivalent promoters, as described previously (Ernst and Kellis 2010, 2012). Notice the presence of two states for which no functional association has been attributed: H3K4me3, or H3K27ac alone. (L) Gene promoters presenting different chromatin co-occurring states across the tissue. (M, top) Local read counts promoter enrichment for the gene Hoxb4 in either of the histone modification maps within the pseudomap. (Bottom) Hoxb4 coenrichment patterns revealing the presence of either bivalent (H3K27me3/H3K27ac or H3K27me3/H3K4me3) or promoter active regions (H3K27ac/H3K4me3). (N) In situ hybridization (ISH) and gene expression data (Allen Mouse Brain Atlas) for the gene Hoxb4, revealing its spatial signature coherent with the H3K27ac digitized view displayed in M.

This Article

  1. Genome Res. 35: 1633-1645

Preprint Server