Why do .wig files differ in lines

[dibrahim]

Hi,

I want to correlate the coverage two ChIP-seq experiments by comparing the coverage of extended reads (to mean fragment length) in non-overlapping intervals.

Therefor, I've used the IGV-tools count option to generate .wig files from my sorted .bam files. Now the output file looks like this

variableStep chrom=chr10 span=100
1 0.68
101 1.0
201 1.56
301 2.08
401 3.42
501 2.06
601 1.45
701 1.0
801 1.26
901 1.85

I would now like to correlate two of these experiements in R BUT the output files of the various experiments don't have the same line numbers. So I cannot generate vectors where the one line matches the other

I don't understand why the count program doesn't generate two files with the same line number? That's what I should expect, right?

Does anyone know why the line number is different? Does anyone have a better way to correlate the datasets in non-overlapping intervals?

Thanks

Daniel

[jujubix]

Without seeing the files or knowing how IGV-tools count works, this is purely a guess, but could it be omitting intervals with 0 coverage, which a variableStep .wig file can do? I recall encountering similar problems matching up two wig files, especially in regions of the chromosomes where coverage is absent.

My solution was to write my own read counting program, or you could identify the missing intervals (e.g. setdiff in R) and manually insert them.

[DZhang]

yes, WIG files omit positions with zero.

Best regards,
Douglas

https://www.contigexpress.com