Fig 1.
Chromatin states of Arabidopsis thaliana.
(A) Bubble plot showing the emission probabilities for histone modifications/variants across the 26 chromatin states of Arabidopsis thaliana. The size of the bubble represents the emission probability ranging from 0 to 1. Coloured rectangles demarcate the classification of states into major domains of chromatin. The box plot on top shows the average ATAC-seq signal for each state representing chromatin accessibility. The two boxes per state are the two replicates of the ATAC-seq experiment. (B) Stacked bar plot showing the overlap between annotated genomic features and chromatin states. (C) Neighbourhood Enrichment analysis of chromatin states representing the fold enrichment for each state at fixed positions relative to the anchor position (TSS and TTS). (D) Metaplot showing the enrichment of prominent histone variants and modifications associated with active transcription. The enrichments here are plotted on genes overlapping states E1-E8. (n = 15234/ 27416 protein coding genes).
Fig 2.
Chromatin states of Marchantia polymorpha.
(A) Bubble plot showing the emission probabilities for histone modifications/variants across the 15 chromatin states of Marchantia polymorpha. The size of the bubble represents the emission probability ranging from 0 to 1. Coloured rectangles demarcate the classification of states into major domains of chromatin. The box plot on top shows the average ATAC-seq signal for each state representing chromatin accessibility. The two boxes per state are the two replicates of the ATAC-seq experiment. (B) Stacked bar plot showing the overlap between annotated genomic features and chromatin states. (C) Neighbourhood Enrichment analysis of chromatin states representing the fold enrichment for each state at fixed positions relative to the anchor position (TSS and TTS). (D) Metaplot showing the enrichment of prominent histone variants and modifications associated with active transcription. The enrichments here are plotted on genes overlapping states MpE1-MpE4. (n = 10715/ 17944 protein coding genes). (E) Enrichment heatmap and metaplots showing the apparent disassociation between H3K27me3 and H2AKub marks in wild type and Mpknox2 Mpe(z)1 double mutants. The enrichment of both marks is plotted on wild type H3K27me3 peaks (n = 25159). The peaks have been clustered using k-means clustering into two clusters where the cluster 1 represents peaks of H3K27me3 shared with H2AKub while cluster 2 represents majority of H3K27me3 peaks that do not overlap with H2AKub.
Fig 3.
Association between chromatin states and transcription across promoters in Arabidopsis and Marchantia.
(A, D) Bar plot of the proportion of chromatin states for 20 bins across promoters of genes ordered by transcripts per million (TPM) Arabidopsis leaves (A) and Marchantia thallus (D). (B, E) Bar plot of the proportion of chromatin states across promoters for 20 bins of genes ordered by coefficient of variation (CV) across different tissues of Arabidopsis (D) or Marchantia (E). (C, F) Bar plot the proportion of chromatin states across promoters for genes all genes or differentially expressed in specific tissues in Arabidopsis (C) and Marchantia (F). (G, H) Bar plot of the proportion of chromatin states across promoters for genes covered by H3K27me3 or differentially expressed in PRC2 mutant in Arabidopsis (G) and Marchantia (H). clf/swn double mutant for Arabidopsis and Mpknox2 Mpe(z)1 double mutant for Marchantia, all genes are shown as a reference in both cases. In the right, asinh (TPM) of expression values for the same group of genes are shown for both mutant and the WT. The color code for the diagram represents the chromatin states in the 15 states models.
Fig 4.
Transcription factors activity shows heterogenous chromatin state preferences in Arabidopsis.
(A-C) Boxplot of TF occupancy scores based on ChIP-seq (A), DAP-seq (B), or position weight matrixes (PWM) motifs (C) in Arabidopsis. (D, E) Comparison of TF occupancy (D) and TF activity (E) scores across different chromatin states. The latter scores are calculated as the statistical enrichment of TF binding over putative targets (co-expressed genes). (F) Heatmap and hierarchical clustering of TF activity scores for ChIP-seq and DAP-seq experiments across chromatin states. Colour scale is shown at the bottom. Scores are normalized to the maximum value of each TF. Low unnormalized scores were discarded. The colour code for the diagrams represents the chromatin states in the current 15 states models. (G) Histogram of TF binding distribution relative to the transcription starting site (TSS) for Arabidopsis TFs from different clusters defined in Fig 4F. (H) Bubble plot of Fisher’s Test of TF families across clusters defined in Fig 4F.
Table 1.
Reagents and tools.
Table 2.
Data to score TF activity.