Cnidarian chromatin accessibility and gene regulation

This repository contains code for downstream analysis of Nemtostella vectensis scATAC-seq data, including clustering and annotation, motif analysis, gene regulatory networks and sequence models.

Table of Contents

• Code
  • Data Processing & Clustering
  • Peak Analysis
  • Motif Analysis
  • Gene Regulatory Networks
  • Advanced Analyses
  • Visualization & Documentation
  • Utility Functions
  • Specialized Functions
  • Interactive Applications
  • Processing Scripts
• Data Files
  • Annotations and Genomes
  • Results
    • Cell Type Annotations
    • Peaks
    • Motifs
    • Mapping of Motifs to TFs
    • Gene Regulatory Networks (GRNs)

Code

The code is organized into numbered notebooks and scripts that follow a logical progression from data processing to downstream analyses.

Data Processing & Clustering

1. 01_ArchR_adult.qmd and 01_ArchR_gastrula.qmd - ArchR project setup for adult and gastrula scATAC-seq data

   • Cell filtering, doublet removal, dimensionality reduction
   • UMAP embedding and clustering using ArchR framework

2. 02_Integration.qmd - Integration of adult and gastrula datasets

   • Joint analysis and harmonization of the two datasets
   • Cell type annotation and comparison across developmental stages

Peak Analysis

3. 03_Peaks.qmd - Joint peak calling and mapping

   • Identification of accessible chromatin regions across datasets
   • Peak-to-gene assignment and regulatory element annotation

4. 04_Metacell_mapping.qmd - Metacell analysis

   • Mapping between scATAC and scRNA-seq metacells

Motif Analysis

5. 05_Archetypes.qmd & 05_Archetypes_JSD.ipynb - Motif archetype analysis

   • Identification of regulatory motif patterns
   • Jensen-Shannon Divergence motif similarity analysis

6. 06_Motif_assignment.qmd - Motif-to-TF assignment

   • Comprehensive motif enrichment analysis
   • Assignment of motifs to transcription factors

Gene Regulatory Networks

7. 07_insilico_ChIP.qmd - In silico ChIP-seq analysis

   • ChromVAR analysis for TF activity inference

8. 08_GRN.qmd - Gene Regulatory Network construction

   • Integration of motif data with expression to build GRNs

Advanced Analyses

9. 09_Cytoscape.ipynb - Network visualization

   • Cytoscape-compatible network export and visualization

10. 10_Modules.qmd - Gene module analysis

    • Identification of co-regulated gene modules

11. 11_gkmSVM.qmd & 11_gkmSVM.ipynb - Machine learning models

    • gkm-SVM classifiers for cell type prediction from sequence
    • Comprehensive Python notebook with 68 cells for model training/evaluation

12. 12_chromBPNet.qmd - chromBPNet models

    • Training bias model and cell type-specific chromBPNet models

13. 13_CREsted_trained.ipynb, 14_CREsted_eval.ipynb, 15_CREsted_explain.ipynb - crested model

    • Training, evaluation, and interpretation of crested models

14. 16_sPyce.ipynb - cross-species scATAC integration

    • Integration with mouse data

Visualization & Documentation

90_Figures.qmd - Figure generation for publication 90_Genome_browser.ipynb - Genome browser track preparation 90_Supplementary.qmd - Supplementary analysis and figures

Utility Functions

• utils.py - Python utilities for k-mer analysis and plotting
• scripts/functions.py - Additional Python helper functions
• scripts/scatac_helper_functions.R - R helper functions
• scripts/chromvar_utils.R - ChromVAR-specific utilities

Specialized Functions

• motif-analysis/ - Motif analysis functions
• metacell_downstream_functions/ - Metacell analysis functions

Interactive Applications

• apps/scatac_atlas/ - Chromatin accessibility atlas app (link)
• apps/motif_syntax/ - Motif co-occurrence visualization app (link)

Processing Scripts

• scripts/ - Various shell and R scripts for:
  • HOMER motif analysis
  • gkmSVM training/prediction
  • STREME motif discovery
  • H3K4me3 signal analysis

Data Files

Annotations and Genomes

The analysis makes use of the following genome and annotation files.

• Genome: genome/Nvec_vc1.1_gDNA.fasta (not on Gihub due to size, available from NCBI)

• Gene models: genome/Nvec_v4_merged_annotation_sort.gtf.gz and genome/Nvec_v4_merged_annotation_sort.bed

• Gene annotations for all genes and transcription factors: annotation/Nematostella_DToL_FINAL.tsv and annotation/Nematostella_DToL_TFs_FINAL.tsv

• GO functional annotations: annotation/Nvec_ensembl.GO.rds and annotation/Nvec_ensembl.GO.csv

• PFAM annotations: annotation/Nvec_long.pep.pfamscan_archs.csv

Results

The analysis generates the following key resources listed below.

Cell Type Annotations

• Cell type annotations and metacell mappings: results/Clustering/Annotation_Adult_Gastrula_SEACell.tsv

• Cell type-aggregated peak accessibility, adult and gastrula quantile normalized: results/Clustering/Sum_Adult_Gastrula_Peaks_cell_type_qnorm.rds

• Cell type-aggregated peak accessibility fold changes , adult and gastrula quantile normalized: results/Clustering/Footprint_Adult_Gastrula_Peaks_cell_type_qnorm.rds

• Cell type aggregated gene accessibility fold change: results/GeneScoreMatrix/Matrix-Gene-Scores-cell-type-FC.rds

• UMAP coordinates for metacells: results/Clustering/SEACells_adult_gastrula_UMAP_FC3_gastrula_FC5_adult_qnorm.tsv

Peaks

• All peaks in adult and gastrula: results/Peaks/Peaks_cell_type_mapped.bed

• Filtered peaks assigned to cell types: results/Peaks/Peaks_cell_type_mapped_cell_type_assignment.tsv.gz

• Filtered peaks assigned to genes: results/Peaks/Peaks_cell_type_mapped_gene_assignment_coaccess.tsv.gz

• Filtered peaks assigned to cell types and genes, and classified as promoters (CP, SP, AP) or non-promoters (NO): results/Peaks/Peaks_cell_type_mapped_cell_type_and_gene_assignment_coaccess.tsv.gz

Motifs

• Archetype motifs PWMs:

  • 1,727 motifs used in the downstream analysis: results/Archetypes/motif-archetypes-PPM-PCC-0.8-IC0.5-5bp-pwms.*
  • 1,292 motifs generated using the more stringent minimum motif length filtering: results/Archetypes/motif-archetypes-PPM-PCC-0.8-IC0.5-8bp-pwms.*

• Dictionary mapping input motifs to archetypes: results/Archetypes/motif-archetypes-PPM-PCC-0.8-IC0.5-5bp.dict

• Archetype motif enrichments in cell types: results/Archetypes/motif-enrichment-cell-type-archetypes-PPM-PCC-0.8-IC0.5-5bp-mona-*

Mapping of Motifs to TFs

• Motifs PWMs: results/Motifs/motifs.meme and results/Motifs/motifs.rds

• Combined assignments of all motifs to TFs (left Euler diagram below): results/Motifs/motif-assignment-combined.tsv.gz

• Selected one motif per TF gene (right Euler diagram below): results/Motifs/motif-assignment-selected.tsv.gz

Gene Regulatory Networks (GRNs)

• Global GRN: results/GRN/global_grn.rds

• GRNs per cell type results/GRN/networks/cell_type and broad cell type: results/GRN/networks/broad_cell_type

  • grn_peaks_(cell_type|broad_cell_type)* - GRNs linking TFs to target peaks
  • grn_genes_(cell_type|broad_cell_type)* - GRNs linking TFs to target genes
  • grn_tfs_(cell_type|broad_cell_type)* - GRNs linking TFs to target TFs
  • grn_tfs_info_* - information about TFs in the GRNs (no links)