Is Bam One-Based Or Zero-Based?

Question

[Alex Reynolds]

Reading section 1.2 of the SAM specification, it says that BAM is 0-based. Yet some folks I know who work with sequencing data in that format say it is 1-based. Who is correct?

Answer 1

[Istvan Albert]

The spec says that the SAM format is one based and that the BAM format is zero based.

But this latter only matters if you access the file directly - if you access a BAM file via a tool like samtools that turns BAM into SAM then it will be turned into a 1 based format.

[Alex Reynolds]

Sneaky! Thanks for the explanation...

[Chirag Nepal]

So far, I though BAM is 0-based. But when I look at BAM description at IGV, it says BAM is 1-based, which is confusing. From ENCODE mapped BAM file, what is the best way to manually confirm if the BAM file is in 0-based or 1-based. For eg: ENCODE CAGE data, is this 0-based or 1-based

# Download Encode BAM
wget
http://hgdownload.cse.ucsc.edu/goldenPath/hg19/encodeDCC/wgEncodeRikenCage/wgEncodeRikenCageHelas3CellPapAlnRep1.bam
wget
http://hgdownload.cse.ucsc.edu/goldenPath/hg19/encodeDCC/wgEncodeRikenCage/wgEncodeRikenCageHelas3CellPapAlnRep2.bam

[Istvan Albert]

It only needs to be treated as zero based if you write a tool that opens the binary BAM file directly and you access and extract the field that contains the coordinate itself. For example you use a programming API in python, java or C. Then you need to regard it as 0 based. In any other interpretation the conversion is being done for you, when IGV shows you a BAM file it has already converted it to SAM format that is 1 based.

The IGV help file is misleading the file is not 1 based, what they show is one based. Similarly when you open a BED file in IGV (BED is also zero based) you will see that it is drawn as a 1 based file. But the file is of course still zero based.

They just convert all data onto the same coordinate system.

Answer 2

[JC]

BAM is 0b, SAM (which is what we can read on the screen) is 1b, maybe that's why the confusion.

Answer 3

[biobot 0.0.77.a.1099]

As you point out, the SAM spec is unambiguous on this; BAM data are 0-based and SAM data are 1-based (sections 1.2 and 3.2). A feature at base 0 in a BAM file will be at base 1 when the data are exported as SAM.