Hi,

I was trying to use phastCons100way.UCSC.hg19 package to retrieve phastCons scores. In the  example given in the reference manual a function is used to calculate the score:

scores(phastCons100way.UCSC.hg19,
       GRanges(seqnames="chr7", IRanges(start=117232380, width=5)))

In this example, is the score returned the average over the 5 positions (width = 5)? Or how is the score to be interpreted?

thanks!

hi,

yes, in this example the score returned is the average of the 5 positions, note that:

scores(phastCons100way.UCSC.hg19, GRanges(seqnames="chr7", IRanges(start=117232380, width=5)))
[1] 0.92
x <- scores(phastCons100way.UCSC.hg19, GRanges(seqnames="chr7",
IRanges(start=117232380:117232384, width=rep(1, 5))))
x
[1] 0.8 0.8 1.0 1.0 1.0
mean(x)
[1] 0.92

cheers,

robert.

I packaged these phastCons conservation scores from the UCSC Genome Browser and AFAIK there is no other annotation package with different phastCons conservation scores. What I could recommend you to try is to download the source tar ball from the annotation package phastCons100way.UCSC.hg19 and look at the file

phastCons100way.UCSC.hg19/inst/extdata/README

It describes how these UCSC data was frozen into the package. You could try to do the same for the phastCons scores of your choice, take a look at the UCSC set of human annotations at:

http://hgdownload.soe.ucsc.edu/downloads.html#human

For instance, the ones derived from the alignment of 46 vertebrate species at:

http://hgdownload.soe.ucsc.edu/goldenPath/hg19/phastCons46way

may lead to less conservative conservation scores. You basically have to replace the phastCons100way.rda file in this 'extdata' directory by the one you build yourself, and adapt the rest of the package changing its name, etc. Note that I'm describing annotations for the human reference genome version hg19, and these annotation sets change for different genome reference versions.

If you do not have experience buildling R packages, take a look at the "Writing R Extensions" manual at:

http://cran.r-project.org/doc/manuals/r-release/R-exts.html

If there would be requests from more users for having packaged the 46way phastcons scores, I would do this myself.

cheers,

robert.

Thanks Robert!

Hmm..looks like the 100say may be too conservative for what I had in mind. Is there a less conservative option available?

Thanks Robert!