Examples

Basic Coverage Analysis

Simple Coverage Track

Generate a basic coverage track from a sorted BAM file:

covtobed sample.bam > coverage.bed

Example output:

chr1    0       1000    0
chr1    1000    1050    5
chr1    1050    1200    12
chr1    1200    1500    8
chr1    1500    2000    0

Quality Control Workflows

Detect Low Coverage Regions

# Find regions with coverage less than 10x
covtobed -x 10 sample.bam > low_coverage.bed

# Merge adjacent low coverage regions
bedtools merge -i low_coverage.bed -d 100 > merged_low_coverage.bed

# Get summary statistics
awk '{sum += $3-$2} END {print "Total low coverage bp:", sum}' merged_low_coverage.bed

Identify High Coverage Regions

# Find regions with coverage greater than 100x (potential PCR duplicates)
covtobed -m 100 sample.bam > high_coverage.bed

# Calculate coverage statistics
awk '{print $4}' high_coverage.bed | sort -n | \
awk '{a[i++]=$1} END {print "Max coverage:", a[i-1]; print "95th percentile:", a[int(i*0.95)]}'

Target Enrichment Analysis

Evaluate Target Coverage

# Generate coverage for target enrichment data
covtobed -q 20 enriched_sample.bam > enriched_coverage.bed

# Find well-covered targets (≥20x)
covtobed -m 20 enriched_sample.bam > well_covered.bed

# Intersect with target regions
bedtools intersect -a targets.bed -b well_covered.bed -wa -wb > covered_targets.bed

# Calculate target coverage statistics
bedtools intersect -a targets.bed -b enriched_coverage.bed -wa -wb | \
awk '{sum+=$7*($6-$5); len+=$6-$5} END {print "Mean target coverage:", sum/len}'

Strand-Specific Analysis

RNA-seq Strand Bias Detection

# Generate strand-specific coverage
covtobed --output-strands rnaseq.bam > stranded_coverage.bed

# Find regions with significant strand bias
awk 'abs($4-$5) > 10 && ($4+$5) > 20' stranded_coverage.bed > strand_biased.bed

# Calculate strand bias ratio
awk '{if($4+$5>0) print $1,$2,$3,$4/($4+$5)}' stranded_coverage.bed > strand_ratios.bed

TradIS/Tn-seq Analysis

# Analyze transposon insertion sites with strand information
covtobed --output-strands tradis.bam > tradis_coverage.bed

# Find upregulated regions (high forward strand coverage)
awk '$4 > 50 && $4 > $5*2' tradis_coverage.bed > upregulated_regions.bed

# Find downregulated regions (high reverse strand coverage)  
awk '$5 > 50 && $5 > $4*2' tradis_coverage.bed > downregulated_regions.bed

Physical Coverage Analysis

Mate-Pair Library Analysis

Physical coverage shows the span between paired reads, useful for assembly validation:

# Calculate physical coverage for mate-pair data
covtobed --physical-coverage mate_pairs.bam > physical_coverage.bed

# Compare with regular coverage
covtobed mate_pairs.bam > regular_coverage.bed

# Find regions with good physical coverage but low read coverage (gaps between mates)
bedtools intersect -a physical_coverage.bed -b regular_coverage.bed -v > mate_pair_gaps.bed
Physical coverage example

Pipeline Integration

Complete Alignment to Coverage Pipeline

#!/bin/bash
# Complete pipeline from FASTQ to coverage analysis

GENOME="reference.fa"
READS1="sample_R1.fastq.gz"
READS2="sample_R2.fastq.gz"
SAMPLE="sample_name"

# Alignment and sorting
bwa mem -t 8 "$GENOME" "$READS1" "$READS2" | \
  samtools view -bS | \
  samtools sort -o "${SAMPLE}.bam"

# Generate coverage with quality filtering
export COVTOBED_QUIET=1
covtobed -q 20 "${SAMPLE}.bam" > "${SAMPLE}_coverage.bed"

# Quality control reports
echo "=== Coverage Statistics for $SAMPLE ==="
awk '{sum+=$4*($3-$2); len+=$3-$2} END {print "Mean coverage:", sum/len}' "${SAMPLE}_coverage.bed"

# Low coverage regions
covtobed -x 10 "${SAMPLE}.bam" | \
  bedtools merge -d 50 > "${SAMPLE}_low_coverage.bed"

echo "Low coverage regions (<10x):"
awk '{sum+=$3-$2} END {print sum " bp"}' "${SAMPLE}_low_coverage.bed"

Multi-Sample Coverage Comparison

#!/bin/bash
# Compare coverage across multiple samples

SAMPLES=("sample1" "sample2" "sample3")

# Generate coverage for all samples
for sample in "${SAMPLES[@]}"; do
    covtobed "${sample}.bam" > "${sample}_coverage.bed"
done

# Create a unified coverage track
bedtools unionbedg -i *_coverage.bed > combined_coverage.bedgraph

# Find consistently low coverage regions across all samples
awk 'NF==6 && $4<10 && $5<10 && $6<10' combined_coverage.bedgraph > consistently_low.bed

# Generate summary report
echo "Coverage comparison summary:"
for sample in "${SAMPLES[@]}"; do
    mean_cov=$(awk '{sum+=$4*($3-$2); len+=$3-$2} END {print sum/len}' "${sample}_coverage.bed")
    echo "$sample: mean coverage = $mean_cov"
done

Advanced Filtering Examples

Version 1.4.0+ Filtering (Default Behavior)

# Default: automatically filters invalid alignments
covtobed sample.bam > filtered_coverage.bed

# Equivalent explicit command
covtobed -d sample.bam > filtered_coverage.bed

# High-quality coverage only
covtobed -q 30 sample.bam > high_quality_coverage.bed

Legacy Behavior (Include All Alignments)

# Include duplicates, failed QC, and non-primary alignments
covtobed --keep-invalid-alignments sample.bam > all_alignments_coverage.bed

# Compare filtered vs unfiltered
echo "Filtered coverage:"
awk '{sum+=$4*($3-$2); len+=$3-$2} END {print "Mean:", sum/len}' filtered_coverage.bed

echo "Unfiltered coverage:"
awk '{sum+=$4*($3-$2); len+=$3-$2} END {print "Mean:", sum/len}' all_alignments_coverage.bed

Assembly Quality Assessment

Metagenomic Assembly Coverage

# Skip small contigs common in fragmented assemblies
covtobed -z 1000 metagenome_reads.bam > assembly_coverage.bed

# Calculate per-contig coverage statistics
awk '{contig=$1; cov+=$4*($3-$2); len+=$3-$2} 
     contig!=last {if(last) print last, cov/len; cov=0; len=0; last=contig}
     END {print contig, cov/len}' assembly_coverage.bed > contig_mean_coverage.txt

# Find well-covered contigs (>10x mean coverage)
awk '$2 > 10' contig_mean_coverage.txt > well_covered_contigs.txt

Output Format Examples

BED Format (Default)

covtobed sample.bam

Output:

chr1    0       100     0
chr1    100     200     15
chr1    200     300     23
chr1    300     400     18

Counts Format

covtobed --format counts sample.bam

Output:

>chr1
0    100
15   100
23   100
18   100

Stranded BED Format

covtobed --output-strands sample.bam

Output (chr, start, end, forward_cov, reverse_cov):

chr1    100     200     8       7
chr1    200     300     12      11
chr1    300     400     10      8

Integration with Other Tools

With bedtools

# Merge adjacent regions with same coverage
covtobed sample.bam | bedtools merge -c 4 -o mean > merged_coverage.bed

# Intersect coverage with gene annotations
bedtools intersect -a genes.bed -b coverage.bed -wa -wb > gene_coverage.bed

# Find coverage gaps in exons
bedtools intersect -a exons.bed -b coverage.bed -v > uncovered_exons.bed

With samtools

# Pre-filter with samtools before coverage analysis
samtools view -q 30 -F 1804 sample.bam | \
  samtools view -bS | \
  covtobed > high_quality_coverage.bed

# Process specific chromosomes
samtools view -b sample.bam chr1 chr2 chr3 | \
  covtobed > main_chromosomes_coverage.bed

With awk for custom analysis

# Calculate coverage distribution
covtobed sample.bam | \
  awk '{print $4}' | \
  sort -n | \
  uniq -c | \
  awk '{print $2, $1}' > coverage_distribution.txt

# Find regions with exactly zero coverage
covtobed sample.bam | \
  awk '$4 == 0' > zero_coverage_regions.bed

# Coverage-weighted GC content (if you have GC content data)
bedtools intersect -a coverage.bed -b gc_content.bed -wa -wb | \
  awk '{sum+=$4*$8*($3-$2); total+=$4*($3-$2)} END {print "GC-weighted by coverage:", sum/total}'