Extended Data Fig. 2: Dynamics of chromatin marks on APREs and reporter assays.

a, Summary of the zebrafish and medaka RNA-seq and ATAC-seq datasets generated for this study. Dashed lines indicate equivalent developmental stages in the two species, based on aprevious study⁷⁰. The number of biological replicates is indicated for each experiment. Zebrafish 24-hpf ATAC-seq data are from a previous study⁶⁷. b, Cumulative distribution of the distance between CAGE-seq peaks and the closest annotated TSSs for genes with expression cRPKM > 5 in any of the samples covered by CAGE-seq (see Fig. 1a). Only CAGE-seq peaks within 1 kbp of an annotated TSS were tested (amphioxus: 10,435 peaks; zebrafish, 23,326 peaks; and mouse, 23,443 peaks). c, Cumulative distribution of distances between each APRE and the closest annotated TSS normalized by the average intergenic distance of the species (amphioxus, 83,471; zebrafish, 252,774; medaka, 174,139; and mouse, 216,857 APREs, as per Fig. 1c). d, Signal distribution of different marks within functional-genomic regions in amphioxus. log₁₀ of read counts of H3K4me3, H3K27ac and ATAC-seq, and raw read counts of CAGE-seq in promoters of homology-supported, protein-coding genes (n = 26,501), other APREs (‘O. APREs’, all APREs that do not overlap a TSS from any gene model; n = 48,341), proximal APREs (n = 24,622), distal APREs (n = 11,881), previously validated enhancers (n = 43; Supplementary Table 9), random regions (n = 88,413) and negative regions (excluding ATAC-seq peaks, n = 88,413). For region designation, see Fig. 1c. For clarity, whiskers and outliers are not displayed. e, k-means clustering of APREs based on H3K27ac signal in three developmental stages. Cluster 1 and 3 APREs were considered as active and inactive, respectively. Average H3K27ac profiles are represented in the top panels. The number of APREs per cluster and stage are provided in Supplementary Data 2, dataset 8. f, Alluvial plot that shows the dynamics of each APRE among the clusters described in e. APREs that remained active (cluster 1 in all stages) along the three developmental stages are represented in blue, constitutively inactive APREs (cluster 3 in all stages) in dark grey and dynamic APREs in red or orange (if inactivated or activated, respectively, during development). Five groups of APREs of special interest are highlighted with stronger colours and named GR1–GR5. g, Representative enriched DNA motifs found in each of the groups described in f. GR1 APREs were enriched in early motifs (for example, Smad3 and Oct4, Sox2 and Nanog); GR3 APREs in motifs of transcription factors involved in the generation of the three germ layers (for example, Foxo3, Sox6 and Sox17); GR4 APREs in tissue-specific transcription factors (for example, Foxa2, Otx2 and Crx); and GR5 APREs in CTCF and CTCF-like (BORIS) motifs. q values as provided by Homer. h, Lateral views of embryos from stable transgenic zebrafish lines at 24 hpf (except for Foxa-243, at 48 hpf) showing GFP expression driven by the amphioxus APREs listed in Supplementary Table 8 and highlighted in Supplementary Fig. 1. The number of independent founders with the same expression were as follows: Six1/2-182 (5/5), Foxa-243 (3/3), Foxa-251 (4/4), FoxC-3067 (6/6) and Pax1/9-157 (3/3). Midbrain expression corresponds to the positive-control enhancer included in the reporter constructs. EN, endoderm; HB, hindbrain; MY, myotomes; PA, pharyngeal arch; SC, spinal cord. Scale bar, 250 μm. i, Lateral views of transient transgenic amphioxus embryos, showing GFP expression driven by the APREs highlighted in Supplementary Fig. 1a, b (Foxa-251 (n = 46 out of 52) and Foxc-3067 (n = 27 out of 35), respectively) and in a previous study⁷¹ (Hox-1655, n = 72 out of 80). For each element, left panels correspond to 3D rendering from sub-stacks and right panels to z-stack sagittal sections. Scale bar, 50 μm. Anterior is to the left and dorsal to the top.