gmane.science.biology.informatics.conductor

Heidi Dvinge | 16 May 19:44

Bug in base.profiles.nc (gcrma)?

Hi,

I'm trying to use gcrma with affinity.source= "local", and a predefined
set of NCprobes with gcrma v. 2.12.0. However, I'm running into problems
with the sub-function base.profiles.nc. I think that where it says:

    if (length(seqs) < length(NCprobe)) {
        cat("\nNote: some of your negative control probes do not have
sequence information\n")
        subIndex2 <- match(c(xy2indices(p$x, p$y, cdf = cdfpackagename),
            xy2indices(p$x, p$y + 1), cdf = cdfpackagename),
            NCprobe)
        subIndex2 <- subIndex2[!is.na(subIndex2)]
        bgy <- bgy[!is.na(subIndex1), ]
    }

isn't is supposed to be:

    if (length(seqs) < length(NCprobe)) {
        cat("\nNote: some of your negative control probes do not have
sequence information\n")
        subIndex2 <- match(c(xy2indices(p$x, p$y, cdf = cdfpackagename),
            xy2indices(p$x, p$y + 1, cdf = cdfpackagename)),       #
moving the ")" here
            NCprobe)
        subIndex2 <- subIndex2[!is.na(subIndex2)]
        bgy <- bgy[!is.na(subIndex1), ]
    }

Thanks

(Continue reading)

headers

Tony Chiang | 16 May 20:09

Re: Bug in base.profiles.nc (gcrma)?

Hi Heidi,

It certainly looks like you found a bug. And I think this has not been fixed
because I just had a peek at the source file. Can you send a note to the
maintainer with your patch? That is can you download the source file via
svn, change the code and then send the maintainer the output of svn diff? If
you don't have svn, let me know and I can do this for you.

It might also be a good idea to update your R since the stable version of R
2.7.0 was released a few weeks ago, and someone might yell at you about
updating your R because there can be instances that your version of some
package is outdated because your R is outdated (not true in this case).

Cheers,
--tony

On Fri, May 16, 2008 at 6:48 PM, Heidi Dvinge <heidi@...> wrote:

> Hi,
>
> I'm trying to use gcrma with affinity.source= "local", and a predefined
> set of NCprobes with gcrma v. 2.12.0. However, I'm running into problems
> with the sub-function base.profiles.nc. I think that where it says:
>
>    if (length(seqs) < length(NCprobe)) {
>        cat("\nNote: some of your negative control probes do not have
> sequence information\n")
>        subIndex2 <- match(c(xy2indices(p$x, p$y, cdf = cdfpackagename),
>            xy2indices(p$x, p$y + 1), cdf = cdfpackagename),
>            NCprobe)

(Continue reading)

headers

Heidi Dvinge | 18 May 20:13

Re: Bug in base.profiles.nc (gcrma)?

Hi Tony,

no svn I'm afraid. If you could do that for me that'd be great. And  
yes, I know I ought to jump to R-2.8 - it's slowly but steadily  
climbing up on my to-do list ;)

Cheers
\Heidi

On 16 May 2008, at 19:09, Tony Chiang wrote:

> Hi Heidi,
>
> It certainly looks like you found a bug. And I think this has not  
> been fixed because I just had a peek at the source file. Can you  
> send a note to the maintainer with your patch? That is can you  
> download the source file via svn, change the code and then send the  
> maintainer the output of svn diff? If you don't have svn, let me  
> know and I can do this for you.
>
> It might also be a good idea to update your R since the stable  
> version of R 2.7.0 was released a few weeks ago, and someone might  
> yell at you about updating your R because there can be instances  
> that your version of some package is outdated because your R is  
> outdated (not true in this case).
>
> Cheers,
> --tony
>
> On Fri, May 16, 2008 at 6:48 PM, Heidi Dvinge <heidi@...> wrote:

(Continue reading)