Input Data

FinaleToolkit works with almost any paired-end sequencing data, in three input formats. Each one must be indexed so the engine can stream fragments efficiently.

Format	Extension	Required index	Reference needed?
BAM	.bam	.bam.bai	No
CRAM	.cram	.cram.crai	Yes (FASTA)
Fragment	.frag.gz or .frag.tsv.bgz	.tbi (Tabix)	No

BAM

A Binary Alignment Map stores sequence-alignment results: where each read maps to the reference, plus information about the read itself. It is not human-readable, but it is space-efficient compared to its plaintext counterpart, SAM.

Note

BAM input must be BAI-indexed. Keep the .bam.bai file alongside the .bam.

CRAM

CRAM is a reference-based binary alignment format. It is smaller than BAM while carrying the same information.

Note

CRAM input must be CRAI-indexed. Keep the .cram.crai file alongside the .cram.

CRAM requires a reference FASTA

Because CRAM encodes reads relative to the reference, you must pass the same FASTA reference used during alignment to any subcommand:

$ finaletoolkit coverage sample.cram intervals.bed -r reference.fa -o cov.bed

The FASTA needs a .fai index, which FinaleToolkit creates automatically if it is missing. A .2bit file cannot be used as the reference when reading CRAM.

Fragment file

A Fragment file (.frag.gz or .frag.tsv.bgz) is derived from a BAM file. It is a block-gzipped BED3+2 file with one row per fragment and these columns:

chrom	start	stop	mapq	strand
contig	0-based start	end	mapping quality	+ or -

Note

Fragment files must be Tabix-indexed. Keep the .tbi file alongside the data.

Example:

#chrom  start   stop    mapq    strand
chr1    10000   10050   60      +
chr1    10050   10100   60      -
chr1    10100   10150   60      +
chr1    10150   10200   60      -
chr1    10200   10250   60      +

Tip

Our FinaleDB database hosts ready-to-use fragment files. Learn more at finaledb.research.cchmc.org.