calico/hendrickson_2017_supplement

Supplemental material for the Hendrickson et. al 2017 paper.

Supplementary code for Hendrickson et al., "Simultaneous profiling of DNA accessibility and gene expression dynamics with ATAC-seq and RNA-seq"

Summary

This is the code for reproducing the bioinformatic analyses of the yeast ATACseq data (section 3.3). We assume that the library was demultiplexed and aligned to the genome as described in sections 3.3.1-3.3.2. Analysis of RNAseq data is described in detail elsewhere.

The processing is done by running atacseq_processing.py as described below. The input is the aligned unsorted BAM file and the output is the bigWig track of normalized insertion densities, python pickle file that contains the insertion coverage for every nucleotide in the genome and tab separated file containing the coverage of insertions in 200 bp bins spanning -1000..1000 bases around the transcription start sites.

Requirements

• python2.7
• numpy
• scipy
• pandas
• pysam (>= 0.18)
• UCSC command line bioinformatic utils (a.k.a. kentUtils)

Installation

Clone git repository

mkdir atacseq_analysis
git clone git@github.com:calico/hendrickson_2017_supplement.git atacseq_analysis

Download example data

cd atacseq_analysis
wget https://hendrickson_2017_supplement.storage.googleapis.com/data.tar.gz
tar xvfz data.tar.gz

Usage

usage: atacseq_processing.py [-h]
                             [--whitelist WHITELIST | --blacklist BLACKLIST]
                             input_file output_dir chrlens_file
                             annotation_file

Example of ATACseq processing

positional arguments:
  input_file            Unsorted BAM file
  output_dir            Destination directory
  chrlens_file          Chromosome length file
  annotation_file       Annotation file

optional arguments:
  -h, --help            show this help message and exit
  --whitelist WHITELIST
                        Name of whitelist file
  --blacklist BLACKLIST
                        Name of blacklist file

Parameter explanation

• input_file - unsorted BAM file with aligned reads
• output_dir - directory to write the output (will be created if missing)
• chrlens_file - TSV of format: chromosome length
• annotation_file - BED or GTF file with a single interval per gene.
• whitelist - if given, the insertion coverage will be normalized so that the total number of insertions in the whitelist is the same between samples.
• blacklist - if given, the insertion coverage will be normalized so that the total number of insertions in the genome outside the blacklist is constant between samples.

Both whitelist and blacklist are TSV files with lines of the format chromosome start end, where the intervals 1-based and closed. Alternatively, also lines with only the chromosome name are also permitted.

Example

cd atacseq_analysis
python atacseq_processing.py data/GCN4_t0_S1_L001_R1_001.mrg.bam processed_data \
                        data/SacCer3.fasta.len data/annotation.bed \
                        --blacklist data/blacklist.tsv