'read' bam file directly

[tyxer]

Hi, there:

I am working on quality statistics on aligned reads of sequencing data.
As all know, bam is the binary format that people can not really 'read' it by lines.
As a result, I have to transform bam file into sam file, which is time and space consuming.
I am wondering are there any tools or libraries by which I can directly work on the bam file (not only for statistics purpose)?
I know there are tools can directly make use of bam files, such as bamtools. But the function of it is not powerful enough for me.
Anyone knows how tools like bamtools make use bam files computationally?

Thank you so much

Yuxiang

[adamdeluca]

That's what unix pipes are for.

samtools view | myprogram

If you want to write a program to directly access bam files, there is a c library (samtools) and a java library (picard).

[tyxer]

Hi, Adamdeluca

OK, thank you. I will try them.

Bests

Yuxiang

[adaptivegenome]

Also bamtools has a nice interface for developers.

[tyxer]

Hi, genericforms:

I did not know bamtools has interface. Do you mean the API?
I only tried the command line functions.

Thank you

Yuxiang

[adaptivegenome]

Hi, genericforms:

I did not know bamtools has interface. Do you mean the API?
I only tried the command line functions.

Thank you

Yuxiang

Bamtools has an exceptional C++ API. Source and docs are here:

https://github.com/pezmaster31/bamtools

You can browse the API commands here in a pretty format:

http://pezmaster31.github.com/bamtools/annotated.html

This is the way to go if you are writing C++ code. You can also use samtools and picard tools for other programming interfaces. Furthermore if you just want to do some quick analysis and not write a complete software package, you can just use samtools or bamtools in command line mode to grab a region of the region (VIEW) and then as mentioned in earlier posts, just pipe to a PERL or PYTHON script.

[tyxer]

Hi, genericforms:

Thank you so much.

It is really helpful.

Bests

Yuxiang