GSE171553 Processing Pipeline

Publication

Nature (2021) — PMID 34819669

Dataset

Mapping cell structure across scales by fusing protein images and interactions

1. 1

   After sequencing, raw reads were aligned to GRCh38 and analyzed following the detailed instructions in ENCODE eCLIP-seq Processing Pipeline v2.2 (https://www.encodeproject.org/pipelines/ENCPL357ADL/).

   eCLIP v2.2 GitHub

   $ Bash example

   # Install STAR (example using conda)
   # conda install -c bioconda star
   
   # Define input files and reference genome index
   READS_R1="raw_reads_R1.fastq.gz"
   READS_R2="raw_reads_R2.fastq.gz"
   GENOME_DIR="/path/to/STAR_index/GRCh38" # Placeholder for GRCh38 STAR index (e.g., from ENCODE or UCSC)
   OUTPUT_PREFIX="aligned_eCLIP_sample"
   
   # Align raw reads to GRCh38 using STAR, following ENCODE eCLIP pipeline recommendations
   STAR \
     --runThreadN 8 \
     --genomeDir "${GENOME_DIR}" \
     --readFilesIn "${READS_R1}" "${READS_R2}" \
     --readFilesCommand zcat \
     --outFileNamePrefix "${OUTPUT_PREFIX}_" \
     --outSAMtype BAM SortedByCoordinate \
     --outSAMattributes All \
     --outFilterMultimapNmax 20 \
     --outFilterMismatchNmax 999 \
     --outFilterMismatchNoverLmax 0.04 \
     --alignIntronMin 20 \
     --alignIntronMax 1000000 \
     --alignMatesGapMax 1000000 \
     --outFilterScoreMinOverLread 0.75 \
     --outFilterMatchNminOverLread 0.75 \
     --limitBAMsortRAM 30000000000
   
   # The ENCODE eCLIP-seq Processing Pipeline v2.2 continues with steps such as:
   # - Adapter trimming and deduplication (often handled by UMI-tools or custom scripts)
   # - Filtering and sorting of BAM files (e.g., using samtools and bedtools)
   # - Peak calling (e.g., using CLIPper: https://github.com/yeolab/clipper)
   # - Control peak calling (e.g., using MACS2 for input controls)
   # - IDR analysis for reproducible peaks (e.g., using merge_peaks: https://github.com/yeolab/merge_peaks)
   # - Generation of bigWig tracks for visualization
   

   Copy View on GitHub
2. 2

   Consistent with the ENCODE standard, reads aligning to artifact-enriched or repetitive genomic regions were removed.

   bedtools (Inferred with models/gemini-2.5-flash) v2.30.0 GitHub

   $ Bash example

   # Install bedtools if not already installed
   # conda install -c bioconda bedtools=2.30.0
   
   # Define input and output file paths
   INPUT_BAM="aligned_reads.bam"
   OUTPUT_BAM="filtered_reads.bam"
   BLACKLIST_BED="GRCh38_unified_blacklist_V2.bed"
   
   # Download ENCODE blacklist file for GRCh38 if not available
   # mkdir -p reference
   # wget -O "${BLACKLIST_BED}.gz" https://raw.githubusercontent.com/ENCODE-DCC/chip-seq-pipeline2/master/references/GRCh38_unified_blacklist_V2.bed.gz
   # gunzip -f "${BLACKLIST_BED}.gz"
   
   # Remove reads aligning to artifact-enriched or repetitive genomic regions using bedtools intersect -v
   bedtools intersect -v -a "${INPUT_BAM}" -b "${BLACKLIST_BED}" > "${OUTPUT_BAM}"

   Copy View on GitHub

Tools Used