Fig. 6

H3K27ac profiling in RA samples reveals disease-specific effects. a Schematic representation of our sample collection design. b UMAP of healthy and RA samples collected from PBMC and synovial fluid. The samples clustered by cell type and by immune context. c Volcano plot showing differentially acetylated (H3K27ac) peaks between RA SF and healthy PBMC monocytes. d, e Examples of unexplained GWAS loci that overlap with regions that with higher H3K27ac activity in RA synovial fluid immune cells. RA risk SNPs were fine-mapped using SuSiE [51] and are shown along with their GWAS − log10p-values. ATAC-seq peaks from Calderon et.al [41] were plotted for comparison. d H3K27ac activity at the FCRL3 promoter is increased in RA SF CD4⁺ T cells (log2CPM: 4.02) compared to RA PBMC CD4⁺ T cells (log2-fold-change: 1.55, log2CPM: 2.46, FDR: 0.016) and healthy PBMC CD4⁺ T cells (log2-fold-change: 1.72, log2CPM: 2.30, FDR: 0.0077). For CD8⁺ T cells, the log2-fold-change is 1.32 compared to healthy PBMC. e H3K27ac activity at the ETV7 promoter is increased in RA SF CD4⁺ T cells (log2CPM: 6.48) compared to RA PBMC CD4⁺ T cells (log2-fold-change: 1.89, log2CPM: 4.60, FDR: 0.0016) and healthy PBMC CD4 T cells (log2-fold-change: 2.47, log2CPM: 4.01, FDR: 5.70×10⁻⁵). For monocytes, the log2-fold-change is 2.02 compared to healthy PBMC. f Forest plot of hetibability enrichment in ATAC-seq peaks (top) and H3K27ac CUT&TAG peaks from various cell types (bottom) computed using stratified LDscore regression [5]. RA heritability enrichments in H3K27ac peaks detected in T cells and B cells from RA synovial fluids are greater than that of ATAC-seq peaks of the same cell types subject to in vitro stimulation. Error bars represent ±1 standard error