GSE249247 Processing Pipeline

Publication

Molecular cell (2024) — PMID 39303722

Dataset

Integrated multi-omics analysis of zinc finger proteins uncovers roles in RNA regulation.

1. 1

   *library strategy: Cut&Run

   ChIP-seq vv2.0.0 GitHub

   $ Bash example

   # Install Nextflow if not already installed
   # conda install -c bioconda nextflow
   
   # Create a placeholder input JSON file for the ENCODE ChIP-seq pipeline.
   # This example assumes paired-end reads for a single ChIP replicate and a single control replicate.
   # Replace 'sample_chip_R1.fastq.gz', 'sample_chip_R2.fastq.gz', etc., with your actual file paths.
   # For Cut&Run, the 'control' might be an IgG sample or a spike-in control, depending on the experimental design.
   cat << EOF > input.json
   {
     "chip": [
       {
         "replicate_name": "chip_rep1",
         "fastq_rep1_R1": "sample_chip_R1.fastq.gz",
         "fastq_rep1_R2": "sample_chip_R2.fastq.gz"
       }
     ],
     "control": [
       {
         "replicate_name": "control_rep1",
         "fastq_rep1_R1": "sample_control_R1.fastq.gz",
         "fastq_rep1_R2": "sample_control_R2.fastq.gz"
       }
     ]
   }
   EOF
   
   # Run the ENCODE DCC ChIP-seq pipeline using Nextflow.
   # The pipeline will automatically handle alignment, peak calling, and quality control.
   # '--genome hg38' specifies the human reference genome (GRCh38/hg38) as a placeholder.
   # You may need to specify a different genome (e.g., mm10 for mouse) or provide a custom genome configuration.
   # '-profile docker' uses Docker containers for reproducibility; ensure Docker is installed and running.
   # '--outdir results' specifies the output directory.
   nextflow run ENCODE-DCC/chip-seq-pipeline2 -profile docker --input input.json --genome hg38 --outdir results

   Copy View on GitHub
2. 2

   Data analysis was performed using a modified version of CUT&RUNTools 2.0.

   CUT&RUNTools v2.0

   $ Bash example

   # Example installation (assuming Python environment and dependencies like numpy, scipy, pysam, deeptools, macs2 are met)
   # git clone https://github.com/yezhengwen/CUT-RUNTools.git
   # cd CUT-RUNTools
   
   # Placeholder for input files and genome reference
   TREATMENT_BAM="path/to/treatment.bam" # Aligned BAM file for treatment sample
   CONTROL_BAM="path/to/control.bam"   # Aligned BAM file for control sample (e.g., IgG or input)
   GENOME_FASTA="path/to/hg38.fa"       # Reference genome FASTA file (e.g., from UCSC or Ensembl)
   OUTPUT_DIR="cutruntools_analysis_output"
   
   # Create output directory if it doesn't exist
   mkdir -p "${OUTPUT_DIR}"
   
   # Execute CUT&RUNTools 2.0
   # Note: The description mentions a "modified version". This command represents the standard usage of CUT&RUNTools 2.0.
   # Specific modifications would need to be applied to the script itself or reflected in additional parameters if applicable.
   python CUT-RUNTools.py \
       -t "${TREATMENT_BAM}" \
       -c "${CONTROL_BAM}" \
       -g "${GENOME_FASTA}" \
       -o "${OUTPUT_DIR}" \
       # Add other parameters as needed, e.g., for species, peak caller, fragment size, etc.
       # -s hg38 \ # Specify species (e.g., hg38, mm10)
       # -p macs2 \ # Specify peak caller (e.g., macs2, seacr)
       # -f 120   # Specify fragment size (if known, otherwise auto-detected)
   

   Copy
3. 3

   Adapters were trimmed using Trimmomatic (v0.36), followed by a second round of trimming to remove any remaining adapter overhang sequences not removed due to fragment read-through.

   Trimmomatic v0.36

   $ Bash example

   # Install Trimmomatic (if not already installed)
   # conda install -c bioconda trimmomatic=0.36
   
   # Define input and output file names
   # Assuming paired-end reads based on "fragment read-through"
   READ1_IN="input_R1.fastq.gz"
   READ2_IN="input_R2.fastq.gz"
   READ1_PAIRED_OUT="output_R1_paired.fastq.gz"
   READ1_UNPAIRED_OUT="output_R1_unpaired.fastq.gz"
   READ2_PAIRED_OUT="output_R2_paired.fastq.gz"
   READ2_UNPAIRED_OUT="output_R2_unpaired.fastq.gz"
   
   # Define adapter file path (replace with actual path to your adapter file)
   # Trimmomatic comes with adapter files in its 'adapters' directory, e.g., TruSeq3-PE.fa
   # The description implies a thorough adapter trimming, including read-through sequences.
   # This is handled by the ILLUMINACLIP step.
   ADAPTER_FILE="/path/to/Trimmomatic-0.36/adapters/TruSeq3-PE.fa"
   
   # Execute Trimmomatic for adapter trimming and basic quality filtering.
   # The ILLUMINACLIP step handles both initial adapter removal and the removal of
   # adapter overhangs due to fragment read-through (especially with the palindrome threshold).
   # Common parameters for ILLUMINACLIP are 2:30:10 (seed mismatches:palindrome_clip_threshold:simple_clip_threshold).
   # Other common quality trimming steps (LEADING, TRAILING, SLIDINGWINDOW, MINLEN) are included as typical usage.
   java -jar /path/to/trimmomatic-0.36.jar PE \
       "${READ1_IN}" "${READ2_IN}" \
       "${READ1_PAIRED_OUT}" "${READ1_UNPAIRED_OUT}" \
       "${READ2_PAIRED_OUT}" "${READ2_UNPAIRED_OUT}" \
       ILLUMINACLIP:"${ADAPTER_FILE}":2:30:10 \
       LEADING:3 \
       TRAILING:3 \
       SLIDINGWINDOW:4:15 \
       MINLEN:36

   Copy
4. 4

   Reads were aligned to hg38 using bowtie2 (v2.3.5.1) using preset “--very-sensitive-local,” minimum fragment length 10, and maximum fragment length 700.

   Bowtie2 v2.3.5 GitHub

   $ Bash example

   # Install bowtie2 (if not already installed)
   # conda install -c bioconda bowtie2
   
   # Install samtools (if not already installed, for converting SAM to BAM)
   # conda install -c bioconda samtools
   
   # Define reference genome index path
   # Replace with the actual path to your hg38 bowtie2 index files
   HG38_INDEX="path/to/hg38_bowtie2_index/hg38"
   
   # Define input FASTQ files (assuming paired-end reads based on fragment length parameters)
   # Replace with your actual input read files
   READS_R1="input_R1.fastq.gz"
   READS_R2="input_R2.fastq.gz"
   
   # Define output BAM file
   OUTPUT_BAM="aligned.bam"
   
   # Align reads to hg38 using bowtie2
   # --very-sensitive-local: preset for sensitive local alignment
   # -I 10: minimum fragment length 10
   # -X 700: maximum fragment length 700
   # -x ${HG38_INDEX}: path to the reference genome index
   # -1 ${READS_R1}: first mate reads
   # -2 ${READS_R2}: second mate reads
   # -S - : output SAM to stdout
   # | samtools view -bS - : pipe SAM from stdout to samtools to convert to BAM
   # > ${OUTPUT_BAM}: redirect samtools output to the BAM file
   bowtie2 --very-sensitive-local -I 10 -X 700 -x "${HG38_INDEX}" -1 "${READS_R1}" -2 "${READS_R2}" -S - | samtools view -bS - > "${OUTPUT_BAM}"

   Copy View on GitHub
5. 5

   After removing PCR duplicates with Picard (v0.1.8), peaks were called using MACS2 (v2.2.7.1) on the default narrowPeak setting using the same-batch V5 mock IP sample as a normalization control.

   MACS2 v2.2.7

   $ Bash example

   # Install MACS2 (if not already installed)
   # conda install -c bioconda macs2
   
   # Define input files and output prefix
   TREATMENT_BAM="treatment.bam" # Replace with your actual treatment BAM file (after PCR duplicate removal)
   CONTROL_BAM="control.bam"     # Replace with your actual control BAM file (V5 mock IP, after PCR duplicate removal)
   OUTPUT_PREFIX="my_experiment_macs2_peaks"
   OUTPUT_DIR="macs2_output"
   
   # Create output directory if it doesn't exist
   mkdir -p "${OUTPUT_DIR}"
   
   # Run MACS2 callpeak with default narrowPeak settings
   # -t: Treatment file (IP sample)
   # -c: Control file (V5 mock IP sample)
   # -f BAM: Input file format (BAM for aligned reads). Assumes single-end reads; use BAMPE for paired-end.
   # -g hs: Effective genome size (e.g., 'hs' for human, 'mm' for mouse). Adjust if needed for your organism.
   # -n: Experiment name, used as prefix for output files
   # --outdir: Specify output directory
   macs2 callpeak \
     -t "${TREATMENT_BAM}" \
     -c "${CONTROL_BAM}" \
     -f BAM \
     -g hs \
     -n "${OUTPUT_PREFIX}" \
     --outdir "${OUTPUT_DIR}"

   Copy

Tools Used