Input Formats¶

BAM files¶

Most tools require a sorted, indexed BAM file as input.

# Sort and index with samtools
samtools sort input.bam -o sorted.bam
samtools index sorted.bam

Note

SAM format is accepted by some tools, but BAM is recommended for performance.

BED12 gene models¶

Many tools require a reference gene model in BED12 format (12-column BED). Each line represents a transcript with exon structure.

The 12 columns are:

Column	Description
chrom	Chromosome
chromStart	Transcript start
chromEnd	Transcript end
name	Transcript name
score	Score (0-1000)
strand	+ or -
thickStart	CDS start
thickEnd	CDS end
itemRgb	Color (usually 0)
blockCount	Number of exons
blockSizes	Comma-separated exon sizes
blockStarts	Comma-separated exon starts (relative to chromStart)

Where to download gene models¶

UCSC Table Browser: genome.ucsc.edu/cgi-bin/hgTables — select "BED" as output format, check "Whole Gene"
GENCODE: Download GTF and convert with gtfToGenePred and genePredToBed from UCSC tools

Tip

Use a comprehensive gene model (e.g., GENCODE basic) that pools all transcripts. This improves junction and exon coverage analysis.

Chromosome size files¶

Required by bam2wig.py. A tab-separated file with chromosome name and size:

chr1    248956422
chr2    242193529
chr3    198295559
...

Generate from a BAM file:

samtools view -H input.bam | grep @SQ | awk '{print $2"\t"$3}' | sed 's/SN://;s/LN://' > chrom.sizes

Or download from UCSC:

fetchChromSizes hg38 > hg38.chrom.sizes

BigWig files¶

Required by geneBody_coverage2.py, normalize_bigwig.py, and overlay_bigwig.py. BigWig is a compressed binary format for continuous signal data. Generate from a wiggle file:

wigToBigWig input.wig chrom.sizes output.bw