Journal: Nucleic Acids Research

Article Title: Boolean logic links chromatin accessibility states to gene expression variability across cell types

doi: 10.1093/nar/gkag230

Figure Lengend Snippet: The ocrRBBR framework for OCR-driven Boolean rule inference explaining gene expression variability. ( A ) In the mouse multiome dataset, nine blood cell lineages—stromal cells, stem cells, DC, myeloid cells, ILC, B, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} $\alpha \beta$\end{document} T, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} $\gamma \delta$\end{document} T, and activated T (ActT) cells—are shown in distinct colors. ( B ) ATAC-seq data are used to identify all OCRs within \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} $\pm$\end{document} 100 kb of gene promoters. ocrRBBR derives Boolean rules among OCRs to explain gene expression variability, as measured by RNA-seq, across 85 cell types spanning nine blood lineages. ( C ) Candidate models are constructed using all combinations of single-, double-, and triple-OCR subsets from the available OCR repertoire (e.g. \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} $\lbrace OCR_A, OCR_B, OCR_C, \dots , OCR_H, OCR_K, OCR_L\rbrace$\end{document} ). ( D ) Each OCR subset is transformed into a set of Boolean rules, which serve as inputs to a ridge regression model used to predict gene expression across cell types. For example, the double-OCR subset \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} $\lbrace OCR_A, OCR_D\rbrace$\end{document} yields rules such as \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} $(OCR_A \wedge OCR_D)$\end{document} , \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} $(\lnot OCR_A \wedge OCR_D)$\end{document} , and so on. Boolean rules receiving positive (orange) coefficients in the fitted model are associated with cell types where the gene is expressed, whereas those with negative (blue) coefficients correspond to cell types with low or no expression. ( E ) Fitted models and their associated Boolean rule sets—such as \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} $(OCR_A \wedge OCR_D) \ \mathrm{or}\ (\lnot OCR_A \wedge OCR_D)$\end{document} , and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} $(OCR_A \wedge OCR_B \wedge OCR_D) \ \mathrm{or}\ (\lnot OCR_A \wedge OCR_B \wedge OCR_D)$\end{document} —are ranked according to their BIC scores. ( F ) Boolean rules are categorized based on the cell types in which they act as active regulators of gene expression.

Article Snippet: Single-cell RNA-seq and ATAC-seq data from the human peripheral blood mononuclear cells (PBMC) granulocyte-sorted 10k dataset (10x Genomics) were processed using Seurat (v5) [ ] and ArchR (v1.0.2) [ ] in R . The raw HDF5 feature-barcode matrix was imported using the Read10X_h5() function, and RNA and ATAC modalities were separated.

Techniques: Gene Expression, RNA Sequencing, Construct, Transformation Assay, Expressing

Journal: Nucleic Acids Research

Article Title: Boolean logic links chromatin accessibility states to gene expression variability across cell types

doi: 10.1093/nar/gkag230

Figure Lengend Snippet: OCR-driven Boolean rules for Spi1 are associated with cell-type-specific chromatin structures involved in its regulation. ( A ) ATAC-seq readouts within \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} $\pm$\end{document} 100 kb of the Spi1 promoter identify candidate OCRs, and RNA-seq quantifies Spi1 expression levels across 85 cell types spanning nine blood lineages. ( B ) A double-OCR subset \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} $\lbrace \mathrm{E}_{-12}, \mathrm{Promoter}\rbrace$\end{document} —including an enhancer located 12 kb upstream of the TSS and the promoter—predicts gene expression variability with the optimal BIC. In the circular heatmap, blood lineages are color-coded similar to panel (A); OCR accessibility, Spi1 expression, and Boolean rule values are shown, with cold colors representing low levels and warm colors high levels. ( C, D ) MSE and Boolean rule coefficients are shown as a function of the regularization parameter \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} $\log (\lambda )$\end{document} in the fitted ridge regression model, where Boolean rules derived from the double-OCR subset are used as predictors of Spi1 expression. ( E ) Two boolean rules \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} $(\mathrm{E}_{-12} \wedge \mathrm{Promoter}) \ \mathrm{or}\ (\lnot \mathrm{E}_{-12} \wedge \mathrm{Promoter})$\end{document} receive positive coefficients, corresponding to chromatin structures associated with gene regulation in myeloid and B cells.

Article Snippet: Single-cell RNA-seq and ATAC-seq data from the human peripheral blood mononuclear cells (PBMC) granulocyte-sorted 10k dataset (10x Genomics) were processed using Seurat (v5) [ ] and ArchR (v1.0.2) [ ] in R . The raw HDF5 feature-barcode matrix was imported using the Read10X_h5() function, and RNA and ATAC modalities were separated.

Techniques: RNA Sequencing, Expressing, Gene Expression, Derivative Assay

Journal: Nucleic Acids Research

Article Title: Boolean logic links chromatin accessibility states to gene expression variability across cell types

doi: 10.1093/nar/gkag230

Figure Lengend Snippet: Cell-type-specific OCR-driven Boolean rules predict gene expression from ATAC-seq data. ( A ) Heatmap of gene scores derived from 1194 cell-type-specific Boolean rules associated with 661 genes. The Boolean rules were selected based on differential gene score patterns across various blood cell lineages, reflecting cell-type-specific regulatory logic. ( B ) Corresponding RNA-seq expression levels for the genes shown in panel a, demonstrating strong concordance between the Boolean rule-derived gene scores and transcriptomic measurements. This comparison highlights the predictive power of the OCR-driven Boolean rules for gene expression across different cell types.

Article Snippet: Single-cell RNA-seq and ATAC-seq data from the human peripheral blood mononuclear cells (PBMC) granulocyte-sorted 10k dataset (10x Genomics) were processed using Seurat (v5) [ ] and ArchR (v1.0.2) [ ] in R . The raw HDF5 feature-barcode matrix was imported using the Read10X_h5() function, and RNA and ATAC modalities were separated.

Techniques: Gene Expression, Derivative Assay, RNA Sequencing, Expressing, Comparison

Journal: bioRxiv

Article Title: Stage-dependent chromatin accessibility remodeling defines an architectural state in facioscapulohumeral muscular dystrophy myoblasts

doi: 10.64898/2026.03.06.710110

Figure Lengend Snippet: (A) Chromosome-wide distribution of differentially accessible regions across all human autosomes. Each horizontal track represents an individual chromosome, arranged from chromosome 1 to chromosome 22. Up-DARs are indicated in red and Down-DARs in teal. Black and gray segments represent cytogenetic bands (cytobands) along each chromosome. The positions of DARs are shown relative to chromosomal coordinates, illustrating their genome-wide distribution. (B) Volcano plot showing differential chromatin accessibility between FSHD and control primary human myoblasts. Each point represents an individual ATAC-seq peak. The x-axis shows the log₂ fold change in accessibility (FSHD/CTRL), and the y-axis shows the −log₁₀(FDR). Regions with significantly increased accessibility in FSHD (Up-DARs) are shown in red, and regions with significantly decreased accessibility (Down-DARs) are shown in teal. Vertical dashed lines indicate fold-change thresholds, and the horizontal dashed line indicates the statistical significance threshold. ( 1 =chr19:42663366-42663803; 2 =chr8:125236306-125237721; 3 =chr4:37517813-37518483; 4 =chr15:84909779-84910176; 5 =chr7:146872319-146873047; 6 =chr1:180272411-18027321; 7 =chr7:41837735-41842398; 8 =chr7:146578703-146579358; 9 =chr7:147089937-147090658; 63012 =chr7:37253326-37254866; 63013 =chr8:11514184-11516273; 63014 =chr5:55485814-55486163; 63015 =chr21:30704676-30707177; 63024 =chr8:11518982-11519321). (C) Volcano plot showing differential chromatin accessibility between FSHD and control myotubes, displayed as in panel B. Each point represents an individual ATAC-seq peak, with Up-DARs shown in red and Down-DARs shown in teal. ( 1 =chr7:76053318-76053881; 2 =chr1:16126339-16127460; 3 =chr10:131229506-131231084; 4 =chr22:21699861-21700117; 5 =chr14:2770953-2771283; 6 =chr9:80192322-80193046; 7 =chr8:11514182-11516352; 8 = chr11:221023-221287; 9 =chr3:69747056-69747802). (D) Genomic annotation of Up-DARs identified in human primary myoblasts (HPMs; total = 5,102). Peaks were classified based on genomic location relative to annotated gene features, including promoter, exon, intron, 5′ untranslated region (UTR), 3′ UTR, downstream region, and distal intergenic region. Percentages and absolute counts are indicated. (E) Genomic annotation of Down-DARs identified in HPMs (total = 7,491), classified according to genomic features as in panel C. (F) Genomic annotation of Up-DARs identified in myotubes (total = 74), categorized according to genomic features as described above. (G) Genomic annotation of Down-DARs identified in myotubes (total = 89), categorized according to genomic features as described above.

Article Snippet: Primary read processing and downstream ATAC-seq data analysis were performed by Macrogen (Seoul, South Korea).

Techniques: Genome Wide, Control

Journal: bioRxiv

Article Title: Stage-dependent chromatin accessibility remodeling defines an architectural state in facioscapulohumeral muscular dystrophy myoblasts

doi: 10.64898/2026.03.06.710110

Figure Lengend Snippet:

Article Snippet: Primary read processing and downstream ATAC-seq data analysis were performed by Macrogen (Seoul, South Korea).

Techniques:

Journal: bioRxiv

Article Title: Stage-dependent chromatin accessibility remodeling defines an architectural state in facioscapulohumeral muscular dystrophy myoblasts

doi: 10.64898/2026.03.06.710110

Figure Lengend Snippet: (A) TSS enrichment profile showing the distribution of ATAC-seq signal centered on annotated transcription start sites (TSS). The x-axis indicates the distance from the TSS (kb), and the y-axis shows the smoothed peak density. Profiles are shown separately for regions with increased accessibility in FSHD (Up-DARs, red) and decreased accessibility in FSHD (Down-DARs, teal). (B) Distribution of differentially accessible regions overlapping promoter regions of protein-coding genes. Promoter-associated DARs were categorized based on their distance from the TSS: proximal promoter (≤1 kb), intermediate promoter (1–2 kb), and distal promoter (2–3 kb). Pie charts show the proportion and absolute counts of Up-DARs (total = 467) and Down-DARs (total = 468) in each promoter interval. (C) Volcano plot showing chromatin accessibility changes at promoter regions of protein-coding genes. Each point represents an individual promoter-associated ATAC-seq peak. The x-axis indicates log₂ fold change (FSHD/CTRL), and the y-axis shows −log₁₀(FDR). Up-DARs are shown in red and Down-DARs in teal. Selected genes are labeled. Bar plots on the left and right show selected Gene Ontology (GO) biological process terms associated with genes linked to Down-DARs and Up-DARs, respectively. The x-axis indicates −log₁₀(adjusted P-value). (D) Representative Integrative Genomics Viewer (IGV) screenshot showing an example of a DOWN-DAR overlapping a promoter region of a protein-coding gene. ATAC-seq signal tracks (scale 0-150) from control and FSHD samples are shown aligned to genomic coordinates. (E) Representative IGV screenshot showing an example of a UP-DAR overlapping a promoter region of a protein-coding gene (scale 0-130), displayed as described in panel D. (F) Distribution of peak width for promoter-associated DARs. Violin plots show peak dimension (kb) for Up-DARs and Down-DARs stratified by promoter distance category (proximal, intermediate, distal) and combined categories. Individual data points and median values are shown. Statistical significance is indicated (****). (G) Motif enrichment analysis for transcription factor binding motifs within promoter regions associated with Down-DARs. Each point represents a transcription factor motif. The x-axis shows log₂ odds ratio, and the y-axis shows −log₁₀(q-value). Selected transcription factors are labeled. (H) Motif enrichment analysis for transcription factor binding motifs within promoter regions associated with Up-DARs. Each point represents a transcription factor motif. The x-axis shows log₂ odds ratio, and the y-axis shows −log₁₀(q-value). Selected transcription factors are labeled.

Article Snippet: Primary read processing and downstream ATAC-seq data analysis were performed by Macrogen (Seoul, South Korea).

Techniques: Labeling, Control, Binding Assay

Journal: bioRxiv

Article Title: Stage-dependent chromatin accessibility remodeling defines an architectural state in facioscapulohumeral muscular dystrophy myoblasts

doi: 10.64898/2026.03.06.710110

Figure Lengend Snippet: (A) Percentage of Down-DARs, Up-DARs, all DARs, and all ATAC-seq peaks overlapping nucleolus-associated domains (NADs). Bar plots show the fraction of regions overlapping NADs. Statistical significance was assessed using Fisher’s exact test relative to all ATAC-seq peaks, as indicated. (B) Chromosome-specific enrichment of DARs within nucleolus-associated domains (NADs). Heatmap showing log₂ odds ratio (OR) values for enrichment of Down-DARs and Up-DARs across chromosomes 1–22. Each row represents a chromosome, and columns indicate Down-DARs and Up-DARs. Color intensity corresponds to log₂(OR), and statistical significance is indicated. (C) Percentage of Down-DARs, Up-DARs, all DARs, and all ATAC-seq peaks overlapping lamina-associated domains (LADs). Bar plots show the fraction of regions overlapping LADs. Statistical significance was assessed using Fisher’s exact test relative to all ATAC-seq peaks, as indicated. (D) Chromosome-specific enrichment of DARs within lamina-associated domains (LADs). Heatmap showing log₂ odds ratio values for enrichment of Down-DARs and Up-DARs across chromosomes. Each row corresponds to a chromosome, and columns indicate Down-DARs and Up-DARs. Color intensity represents log₂(OR), and statistical significance is indicated. (E) Compartment enrichment analysis showing log₂ odds ratio values for enrichment of Down-DARs, Up-DARs, and DAR hotspots in NAD and LAD compartments. Values are shown together with corresponding confidence intervals. (F) Chromatin state composition of DARs. Stacked bar plots showing the fraction of Down-DARs and Up-DARs assigned to different chromatin states, including heterochromatin, weak transcription, weak enhancer, transcription elongation, strong enhancer, repressed, transcription transition, insulator, weak promoter, active promoter, poised promoter, and repetitive/CNV. (G) Chromatin state distribution of DARs across chromosomes. Bubble plots showing the percentage of DARs associated with each chromatin state across chromosomes 1–22 and chromosome X. Circle size represents the fraction of DARs, and color indicates log₂ odds ratio for enrichment. Separate panels are shown for Down-DARs and Up-DARs. (H) Enrichment of repeat element classes in Up-DARs. Bar plot showing log₂ odds ratio values for overlap between Up-DARs and annotated repeat classes, including LINE, SINE, DNA elements, simple repeats, and LTR elements. (I) Enrichment of repeat element classes in Down-DARs. Bar plot showing log₂ odds ratio values for overlap between Down-DARs and annotated repeat classes, including LINE, SINE, simple repeats, low complexity regions, and tRNA elements.

Article Snippet: Primary read processing and downstream ATAC-seq data analysis were performed by Macrogen (Seoul, South Korea).

Techniques:

Journal: bioRxiv

Article Title: MNS induces antiviral protection and suppresses inflammation

doi: 10.64898/2026.02.11.705318

Figure Lengend Snippet: (A) Heatmap of inflammatory pathway gene expression in cells treated with DMSO, LPS, LPS+MNS, or MNS alone, highlighting MNS-mediated regulation relative to LPS. (B) Volcano plot showed DEGs between the treatment of LPS+MNS and LPS alone in RAW264.7 cells. (C) Gene Ontology (GO) enrichment of DEGs in (B). (D) KEGG enrichment of DEGs in (B). (E) Heatmap showed that MNS markedly changed the chromatin accessibility induced by LPS. (F) Heatmap illustrated the differentially accessible regions (DARs), following treatment with DMSO, LPS, LPS-MNS and MNS alone. (G) Distribution of peaks and transcription factor-binding loci relative to transcription start sites (TSS) in the LPS, LPS-MNS, and MNS groups compared to the DMSO group, respectively, as annotated using the ChIPseeker R package. (H) Scatter plot analysis comparing fold changes from ATAC-seq and RNA-seq. (I) Overlap downregulated genes in ATAC-seq and bulk RNA-seq between LPS and LPS-MNS groups in RAW 264.7 cells. (J) KEGG enrichment of overlap genes in (I).

Article Snippet: For ATAC-seq high-throughput data generated by GENEWIZ (Suzhou, China), raw FASTQ files were first subjected to quality control and adapter trimming with Trim Galore (v0.6.4).

Techniques: Gene Expression, Binding Assay, RNA Sequencing