summary bigWig rtracklayer

[klebsiella]

Hi Guys,
The below function in R crashes if no score exists for one of the intervals in the bigwig file. This is annoying as the whole run is stopped instead of issuing a warning.
For Het chromosomes, I tried to escape them, as you can see below. But for other cases, this cannot be generalized

Do you know a way to get around ?


"x" is a data frame of three columns

getBigWigScore<- function(x, bwf, type="mean"){
require(rtracklayer)
bwf <- BigWigFile(bwf)
features <- GRanges(x[,1], IRanges(x[,2], x[,3]))
## the summary below dies if a feature has no score found in the bw file instead of throwing a warning or return NA ...
features <- features[!seqnames(features) %in% c("chr2LHet", "chr2RHet", "chr3LHet", "chr3RHet", "chrXHet", "chrYHet", "chrUextra")]
return(unlist(runValue(summary(bwf, features, type=type))))
}

Best and many thanks for any help,

[dpryan]

I'm not even sure how you could have a valid bigWig (or even Wig) file without all of the positions having scores. It would seem to make sense to just remove those regions by adding multiple "definition" lines.

[klebsiella]

Thanks for the reply,
Well I validated the bigwig files and they contain no errors.
For instance, the problem occurred for a feature starting with "0" (0 based coordinates file).
I agree with you, a wig file is supposed to contain continuous data.
Also, could you please elaborate on what you mean by adding a few definition lines ?

A work around for me now was to use either Java Genomics Toolkit or bwtool (very very fast) via the system call from R.
Thanks

[dpryan]

Sorry for the delay, I just realized that I never replied to this!

You can have multiple definition lines like the following the remove the need for having lines with no values:

fixedStep chrom=chr1 start=200601 step=100
11
22
33
fixedStep chrom=chr1 start=401601 step=100
111
122
133

That way, you can just leave out whatever portions aren't covered.