Bio Methylation Bismark Alignment Agent Skill

# One-time genome preparation (creates bisulfite-converted index)
bismark_genome_preparation --bowtie2 /path/to/genome_folder/

# Genome folder should contain FASTA files (e.g., hg38.fa, chr1.fa, etc.)
# Creates Bisulfite_Genome/ subdirectory with CT and GA converted indices

bismark --genome /path/to/genome_folder/ reads.fastq.gz -o output_dir/

bismark --genome /path/to/genome_folder/ \
    -1 reads_R1.fastq.gz \
    -2 reads_R2.fastq.gz \
    -o output_dir/

bismark --genome /path/to/genome_folder/ \
    --bowtie2 \                    # Use bowtie2 (default)
    --parallel 4 \                 # Number of parallel instances
    --temp_dir /tmp/ \             # Temporary directory
    --non_directional \            # For non-directional libraries
    --nucleotide_coverage \        # Generate nucleotide coverage report
    -o output_dir/ \
    reads.fastq.gz

# Reduced Representation Bisulfite Sequencing
bismark --genome /path/to/genome_folder/ \
    --pbat \                       # For PBAT libraries (post-bisulfite adapter tagging)
    reads.fastq.gz

# MspI digestion (RRBS standard)
# Bismark handles MspI-digested libraries automatically

# Post-Bisulfite Adapter Tagging (e.g., scBS-seq)
bismark --genome /path/to/genome_folder/ --pbat reads.fastq.gz

# For libraries where all 4 strands are present
bismark --genome /path/to/genome_folder/ --non_directional reads.fastq.gz

# Trim adapters first with Trim Galore (recommended)
trim_galore --illumina --paired reads_R1.fastq.gz reads_R2.fastq.gz

# Then align
bismark --genome /path/to/genome_folder/ \
    -1 reads_R1_val_1.fq.gz \
    -2 reads_R2_val_2.fq.gz

# --parallel sets instances per alignment direction
# Total threads = parallel * 2 (for directional) or parallel * 4 (non-directional)
bismark --genome /path/to/genome_folder/ \
    --parallel 4 \
    reads.fastq.gz

# Bismark produces:
# - reads_bismark_bt2.bam          # Aligned reads
# - reads_bismark_bt2_SE_report.txt # Alignment report

# View alignment report
cat output_dir/reads_bismark_bt2_SE_report.txt

# Bismark output is unsorted
samtools sort output.bam -o output.sorted.bam
samtools index output.sorted.bam

# Remove PCR duplicates (recommended for WGBS, not RRBS)
deduplicate_bismark --bam output_bismark_bt2.bam

# For paired-end
deduplicate_bismark --paired --bam output_bismark_bt2_pe.bam

# Bismark generates detailed report
cat *_SE_report.txt

# Key metrics:
# - Sequences analyzed
# - Unique alignments
# - Mapping efficiency
# - C methylated in CpG context

# HISAT2 is faster and uses less memory for large mammalian genomes
bismark_genome_preparation --hisat2 /path/to/genome_folder/

# Align with HISAT2
bismark --genome /path/to/genome_folder/ --hisat2 reads.fastq.gz

# HISAT2 paired-end
bismark --genome /path/to/genome_folder/ --hisat2 \
    -1 reads_R1.fastq.gz \
    -2 reads_R2.fastq.gz