creating a bw file about pybigwig

[dpryan79]

While you can make pyBigWig accept input like that, the resulting file will largely be useless. For a bigWig file to have normal functionality, all entries for each chromosome must be together. This is true regardless of the program used.

Sent from my iPhone

On 21 May 2016, at 14:37, PanosFirmpas [email protected] wrote:

hello,

I'm writing a script that goes from a bam input to a bw output (it's in my github page if you feel like checking it out, input would be very welcome!).

Right now, my script outputs wiggle "strings" in a subprocess pipe to wigToBigWig(the ucsc utility). This is very convenient since wigtobigwig seems to accept the input with no concerns that the chromosomes are out of order, as long as the input per chromosome is in order

for example:

chromosome 2 start:100 step 1:
1
2
1
chromosome 1 start:100 step 1:
1
2
1
chromosome 2 start:200 step 1:
1
2
1
chromosome 1 start:200 step 1:
1
2
1

works fine because the input for each chromosome is in proper order.

does that work with pyBigWig?

If it does i think it should provide a nice speed boost for my script !

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub

[PanosFirmpas]

Could you please elaborate a bit on the file being useless? the output i get seems like what you'd expect it to be (on the browser, around genomic locations we're familiar with), am I getting noise and haven't noticed?

[dpryan79]

Presumably the program you're feeding interleaved regions to sorts them before actually writing them to a file. bigWig files themselves actually require that intervals from a given chromosome be next to each other and in a sorted order in order for the indexing to work. Since pyBigWig doesn't sort the regions for you it'll produce the file with regions in the order provided...which will produce a bigWig file with a broken index.

Anyway, unless validate=False is used, pyBigWig will also prevent creation of a file with a broken index.

[dpryan79]

As an aside, your bamToBigWig program sounds like our bamCoverage program from deepTools.

[PanosFirmpas]

It does look very similar indeed, I'm sorry i didn't know about your tool, but well this was a nice learning experience for me !

Thank you for the comments on bigwigs, it looks like i need to do some testing!!