This repository contains analysis code associated with Ing-Simmons et al., Nature Genetics, 2021, 'Independence of chromatin conformation and gene regulation during Drosophila dorsoventral patterning'.
This code is provided as an enhancement to the Methods described in the paper, in order to increase the reproducibility of our work and help interested readers understand how the analysis was carried out. Please note that while we aim for easy reproducibility, this repository contains the actual code that was developed over time and used for the analysis and may not be a fully self-contained workflow to reproduce all results in the paper.
The analysis is set up as several separate Snakemake workflows, for the initial ChIP-seq analysis, for the scRNA-seq analysis, and one for all other analysis and integration of different datasets.
This analysis uses data from the following ArrayExpress accessions:
- ChIP-seq: E-MTAB-9303
- scRNA-seq: E-MTAB-9304
- Hi-C: E-MTAB-9306
- Micro-C: E-MTAB-9784
We also used publicly available ChIP-seq and RNA-seq data from Koenecke et al. 2016, with GEO accession GSE68983, as well as ChIP-seq data from modENCODE.
Python packages and versions used can be found in requirements.txt. R packages and versions used can be found in renv.lock (for more information on using renv, see here).
This can be found in the external_data/ directory. Each individual data source has its own subdirectory with its own Snakefile. These are set up so that the relationship between files and samples is specified in a metadata.txt file. Only the input fastq files and the metadata.txt files should be needed to run the workflow to produce .bam and .bw files.
This analysis can be found in the scrna-seq/ directory. The data was analysed using Cell Ranger as shown in the Snakefile, with downstream analysis and visualisation in R as shown in the .Rmd file in the scripts/ directory.
This analysis is split over multiple .smk files for better organisation. These can be found in the workflow/ directory. This workflow includes identification of putative tissue-specific enhancers and their downstream analysis, as well as all Hi-C and Micro-C analysis, integrative analysis, and visualisation.
Hi-C analysis was carried out using FAN-C. Both Python and R were used for analysis; individual analysis and plotting scripts used in different steps of the workflow can be found in the scripts/ directory.
The workflow assumes that .fastq files for Hi-C and Micro-C are present in data/fastq/. The relationship between samples and files is described in metadata.txt (and micro-c_metadata.txt).