CDF & CEL file formats

[Frank Ziegner]

Hi everyone,

does anybody know, where to find detailed information about the Affymetrix CDF and CEL file format?

I'm less interested in how the probes for the chip (I work with HuEx-1_0-st-v2.) were selected, this is almost completely covered in this Technote. I'm more interested in questions like how the features were distributed on the chip or if there are calibration markings on the chip for the scanner. Maybe then I could explain some pattern in my spatial plots of the raw data or the residuals. Attached is a spatial plot of the residuals. As discussed some time ago, this plot is not very meaningful since only a fraction of probesets were fitted. But nevertheless it shouldn't show this sort of vertical separation. Also I would like to know, what the blank spots are for.

Also can anybody recommend some literature which describes CDF or CEL files in a more general or introductive way? This would be for an introduction of "newbies" to this topic.

Thanks,
Frank

[Elizabeth Purdom]

Hi Frank,

As I mentioned when I responded to your previous post a while back, I
also get similar 'blocking' patterns in my values (you can see it in the
intensities, not just the residuals). I contacted Alan Williams at
Affymetrix who was involved with the design of the array; he had a
couple if ideas, but has not tried to determine whether they are truly
the cause:

> First, the probes on are arrays are "edge optimized". This means that
> probes are placed on the array such that differences in synthesis steps
> between adjacent features is minimized. As a result probes with similar
> composition end up near each other.
>
> Another possible cause is that we have a number of control type
> probesets (ie antigenomic background controls) as well as a lot of
> development content (ie yeast probesets), the later not being exposed in
> the library files. The layout of this content could explain the patterns
> as well as some controls are only laid out in certain subgrids.

I have not checked with the control probes to see where they lie and if
this explains it (it's a large swath of the array to explain by controls
and development content, though there are a lot of controls). Basically,
I have also not tried to follow up on this.

I found the following from Affymetrix useful for knowing what is
actually contained in different formats (it's a download, each file in
folder is a file type):
https://www.affymetrix.com/support/file_download.affx?onloadforward=/support/developer/fusion/file_formats.zip
It's best when used while also looking at the file (ASCII version).

Best,
Elizabeth

Frank Ziegner wrote:
> Hi everyone,
>
> does anybody know, where to find detailed information about the
> Affymetrix CDF and CEL file format?
>
> I'm less interested in how the probes for the chip (I work with
> HuEx-1_0-st-v2.) were selected, this is almost completely covered in
> this Technote

> <http://www.affymetrix.com/support/technical/technotes/exon_array_design_technote.pdf>.

> I'm more interested in questions like how the features were distributed
> on the chip or if there are calibration markings on the chip for the
> scanner. Maybe then I could explain some pattern in my spatial plots of
> the raw data or the residuals. Attached is a spatial plot of the
> residuals. As discussed some time ago, this plot is not very meaningful
> since only a fraction of probesets were fitted. But nevertheless it
> shouldn't show this sort of vertical separation. Also I would like to
> know, what the blank spots are for.
>
> Also can anybody recommend some literature which describes CDF or CEL
> files in a more general or introductive way? This would be for an
> introduction of "newbies" to this topic.
>
> Thanks,
> Frank
>
> >
>

> ------------------------------------------------------------------------
>