rat_singlecell_liver_ArchR

Branch version1 is the code for the article "Gene-regulation modules in nonalcoholic fatty liver disease revealed by single-nucleus ATAC-seq."

Files are in R Markdown/Notebook (.Rmd) format. They are named 1_, 2_, ... in the order of execution. I recommend using RStudio for editing and running.

Questions? Please submit to GitHub Issues or e-mail fumihiko AT takeuchi DOT name

Data download

Animal ID and experimental condition

• See fields sample_name, age and treatment for each biosample in DDBJ (BioProject PRJDB13870)

Use case 1: Raw data. Not recommended because the data cleaning takes a few days.

• single-nucleus ATAC-seq
  • BAM files produced by Cell Ranger are available from DDBJ (DRA DRA014511 within BioProject PRJDB13870)
  • Start by running 1_unify_multiple_barcodes_for_single_droplet.Rmd
• bulk ATAC-seq
  • FASTQ files are available from DDBJ (DRA DRA014458 within BioProject PRJDB13870)
  • I used the ATAC-seq data processing workflow by Reske et al.

Use case 2: Datasets after cleaning

• single-nucleus ATAC-seq
  • ArchR project files (*.zip; unzip before use) are available from figshare
  • Start by running 2_ArchR_setting.Rmd and then 4_LSI_and_clustering_of_tile_matrix.Rmd
• bulk ATAC-seq
  • BAM files (*.bam *.bai) are available from figshare
  • The data is analyzed in 6_cell_type_composition.Rmd

Use case 3: Generate tables & figures in manuscript

• Run Use case 2.
• 5_LSI_and_clustering_of_peak_matrix.Rmd > Clustering with taylored parameter > Uniform Manifold Approximation and Projection (UMAP)
  • Fig.1B
• 6_cell_type_composition.Rmd > Refine and inspect cell type composition of bulk samples > Cross check, bulk vs snATAC-seq
  • Fig.1D
• 6_cell_type_composition.Rmd > Refine and inspect cell type composition of bulk samples > Test difference and plot cell type composition
  • Fig.1E
• 7_GeneScoreMatrix.Rmd > QC
  • Fig.1C
• 7_GeneScoreMatrix.Rmd > Compare between conditions
  • Table S2, Dataset S1
• 7_GeneScoreMatrix.Rmd > Cross-check with bulk differential gene expression experiments
  • Fig.S1
• 9_GSEA.Rmd > Summarize results across tgt/bgd combinations
  • Fig.2 (The R code outputs result for each gene set. The heatmap was drawn using Excel.)
• 10_GRN.Rmd > Multiple motifs-to-one gene analysis > Nonnegative matrix factorization of regulator-regulatee matrix
  • Fig.3A, Fig.4A, Fig.5A, Fig.6A
• 10_GRN.Rmd > Multiple motifs-to-one gene analysis > Hyperparameters for the discovery of factors (modules) of TF regulation
  • Fig.S3
• 10_GRN.Rmd > Multiple motifs-to-one gene analysis > Inspect enriched gene sets
  • Fig.3B, Fig.4B, Fig.5B, Fig.6B
• 10_GRN.Rmd > Activity score and co-regulation of a gene set > Compute gene set activity score
  • Fig.7 (top panel), Fig.S2 (top panel)
• 10_GRN.Rmd > Activity score and co-regulation of a gene set > Core genes as those central in co-regulation and protein-protein interaction
  • Fig.7 (middle & bottom panels), Fig.S2 (middle & bottom panels)

Information

Please cite

Takeuchi, F. et al., Gene-regulation modules in nonalcoholic fatty liver disease revealed by single-nucleus ATAC-seq, DOI: 10.26508/lsa.202301988

Takeuchi, F. et al., Single-nucleus ATAC-seq elucidates major modules of gene regulation in the development of non-alcoholic fatty liver disease, DOI: 10.1101/2022.07.12.499681 (old manuscripts)