BED file format fundamentals, creation, validation, and basic operations. Covers BED3 through BED12 formats, coordinate systems, sorting, and format conversion using bedtools and pybedtools. Use when working with genomic coordinates or preparing interval files for downstream tools.
BED file format fundamentals, creation, validation, and basic operations. Covers BED3 through BED12 formats, coordinate systems, sorting, and format conversion using bedtools and pybedtools. Use when working with genomic coordinates or preparing interval files for downstream tools.
tool_type
mixed
primary_tool
bedtools
BED File Basics
BED (Browser Extensible Data) format stores genomic intervals. Uses 0-based, half-open coordinates.
BED Format Columns
BED3: chr start end
BED4: chr start end name
BED5: chr start end name score
BED6: chr start end name score strand
BED12: chr start end name score strand thickStart thickEnd rgb blockCount blockSizes blockStarts
Coordinate System
BED uses 0-based, half-open coordinates:
Start: 0-based (first base is 0)
End: exclusive (not included)
Position 100-200 means bases at positions 100-199
Create BED Files
From Text (CLI)
# Create simple BED3echo -e "chr1\t100\t200\nchr1\t300\t400" > regions.bed
# Create BED6 with name and strandecho -e "chr1\t100\t200\tpeak1\t100\t+" > peaks.bed
From Python
import pybedtools
# From string
bed_string = '''chr1\t100\t200\tpeak1\t100\t+
chr1\t300\t400\tpeak2\t200\t-'''
bed = pybedtools.BedTool(bed_string, from_string=True)
# From list of tuples
intervals = [
('chr1', 100, 200, 'peak1', 100, '+'),
('chr1', 300, 400, 'peak2', 200, '-'),
]
bed = pybedtools.BedTool(intervals)
# From pandas DataFrameimport pandas as pd
df = pd.DataFrame({
'chrom': ['chr1', 'chr1'],
'start': [100, 300],
'end': [200, 400],
'name': ['peak1', 'peak2'],
'score': [100, 200],
'strand': ['+', '-']
})
bed = pybedtools.BedTool.from_dataframe(df)
# Save to file
bed.saveas('output.bed')
Sort BED Files
CLI
# Sort by chromosome and positionsort -k1,1 -k2,2n input.bed > sorted.bed
# Using bedtools
bedtools sort -i input.bed > sorted.bed
# Sort by chromosome, start, then endsort -k1,1 -k2,2n -k3,3n input.bed > sorted.bed
Python
import pybedtools
bed = pybedtools.BedTool('input.bed')
sorted_bed = bed.sort()
sorted_bed.saveas('sorted.bed')
import pybedtools
bed = pybedtools.BedTool('input.bed')
# Filter by chromosome
chr1 = bed.filter(lambda x: x.chrom == 'chr1')
# Filter by size
large = bed.filter(lambda x: len(x) >= 100)
# Filter by strand
plus_strand = bed.filter(lambda x: x.strand == '+')
# Filter by score
high_score = bed.filter(lambda x: float(x.score) >= 500)
# Chain filters
result = bed.filter(lambda x: x.chrom == 'chr1'andlen(x) >= 100)
result.saveas('filtered.bed')
Convert Formats
BED to Other Formats
# BED to GFF
bedtools bed12togff -i input.bed > output.gff
# BED to FASTA (extract sequences)
bedtools getfasta -fi reference.fa -bed input.bed -fo output.fa
# BED to FASTA with names
bedtools getfasta -fi reference.fa -bed input.bed -name -fo output.fa
From VCF to BED
# Extract variant positions
bcftools query -f '%CHROM\t%POS0\t%END\n' input.vcf > variants.bed
# Or using awk (simpler for SNPs)
grep -v "^#" input.vcf | awk '{print $1"\t"$2-1"\t"$2}' > snps.bed
From BAM to BED
# Convert alignments to BED
bedtools bamtobed -i input.bam > alignments.bed
# BED12 for spliced alignments
bedtools bamtobed -i input.bam -split > spliced.bed
# Fixed-size windows across genome
bedtools makewindows -g genome.txt -w 10000 > windows_10kb.bed
# Windows with step size (sliding windows)
bedtools makewindows -g genome.txt -w 10000 -s 5000 > sliding_10kb.bed
# Fixed number of windows per chromosome
bedtools makewindows -g genome.txt -n 100 > 100_windows_per_chr.bed
# Windows within BED regions
bedtools makewindows -b regions.bed -w 1000 > windows_in_regions.bed
# Add window ID
bedtools makewindows -g genome.txt -w 10000 -i winnum > numbered_windows.bed
# Source chromosome in name
bedtools makewindows -g genome.txt -w 10000 -i srcwinnum > windows_with_source.bed
Python
import pybedtools
# From genome file
windows = pybedtools.BedTool().window_maker(g='genome.txt', w=10000)
# Sliding windows
windows = pybedtools.BedTool().window_maker(g='genome.txt', w=10000, s=5000)
# From BED regions
bed = pybedtools.BedTool('regions.bed')
windows = pybedtools.BedTool().window_maker(b=bed.fn, w=1000)
windows.saveas('windows.bed')
