[BioC] beadarray package: problem with bead level data

[Krys Kelly]

Dear Jose

As you say, the jpegs are only 8 bit.  Also, Matt Ritchie told me the tiffs
are 80MB and my files are only 880KB.  I've talked to the people who did the
scanning and the tiffs were deleted because they are so big.  However, they
are rescanning them for me, so I should have the tiffs next week.

Thanks for your help.

Krys


-----Original Message-----
From: J.delasHeras at ed.ac.uk [mailto:J.delasHeras at ed.ac.uk] 
Sent: 26 July 2007 17:00
To: Krys Kelly
Subject: RE: [BioC] beadarray package: problem with bead level data


Hi Krys,

Check first that the jpegs are indeed 8-bit (which is not suitable for  
quantitation). Just open one of the jpegs in Photoshop, and go to the  
menu: Image, then Mode. It'll tell you there if it's 8 or 16 bit. If  
it's 16 bit (unlikely), then the problem is something else. If it's  
8-bit as I suspect, then just wait for the tiff files (they should  
have them already) and hopefully they'll look good.

Usually scanning is done at 16-bit, to produce a TIFF file, but jpegs  
may also be generated, at 8-bit, just for illustration purposes. They  
look good to the eye, and being a lot smaller they're better to just  
show what the array looks like, but for quantitation you need the  
16-bit file (higher resolution, if you like). They may have sent you  
the jpegs not knowing you were going to segment/quantitate it yourself.

Again, I must insist, I am not familiar with beadarray and the process  
you were following, so I hope I understood correctly what you were  
trying to do! :-) But the one thing I know is that an 8-bit image will  
not do for quantitation of microarrays.

Good luck,

Jose


Quoting Krys Kelly <kak28 at cam.ac.uk>:

> Hi Jose
>
> Thanks for this.  I have contacted the person who ran the arrays to ask if
> they can scan them again to produce .tif files.
>
> Krys
>
> -----Original Message-----
> From: bioconductor-bounces at stat.math.ethz.ch
> [mailto:bioconductor-bounces at stat.math.ethz.ch] On Behalf Of
> J.delasHeras at ed.ac.uk
> Sent: 26 July 2007 15:29
> To: bioconductor at stat.math.ethz.ch
> Subject: Re: [BioC] beadarray package: problem with bead level data
>
> Quoting Krys Kelly <kak28 at cam.ac.uk>:
>
>>
>> I am having a problem with bead level data.
>>
>> I have .txt files containing over 900,000 lines with four columns: Code,
>> Grn, GrnX and GrnY.
>>
>> The image files are .jpg rather than .tif. I have converted the .jpg
> images
>> to .tif using Adobe Photoshop (trying several of the conversion options).
>> They always look like a beautiful starry night sky, which I think is what
> is
>> expected.
>>
>> However, the data in the beadarray object is not what I expected.  In G
>> there are a lot of negative and zero values and all the Gb values are
> zero.
>>
>> On Wednesday, I downloaded the latest version of R, beadarray and the
>> packages it depends on (see sessionInfo below).
>>
>> Is the problem with my .tif files, or is there a problem of
> incompatibility
>> with the new version of R and the dependencies, or something else?
>>
>> Thanks for any light you can shed.
>>
>> Krys
>
>
> Just a quick comment...
>
> you are converting a jpeg to a tiff, for your quantitation, right?
>
> is the jpeg 16-bit?
>
> Jpegs are *usually* 8-bit, and I'd be surprised if yours is 16-bit,
> but they can be. If it's 8-bit, you're only getting 256 different
> intensities (2^8) rather than 65536 (2^16)... so I wonder if this may
> explain what you're seeing (you'd essentially get what looks like an
> extremely weak hyb, if the software reading the raw data expects a
> range between 1-65536 but the maximum is 256).
>
> I am not familiar at all with beadarray, but your conversion concerns
> me. You'd better check that first.
>
> Jose
>
>
> --
> Dr. Jose I. de las Heras                      Email: J.delasHeras at ed.ac.uk
> The Wellcome Trust Centre for Cell Biology    Phone: +44 (0)131 6513374
> Institute for Cell & Molecular Biology        Fax:   +44 (0)131 6507360
> Swann Building, Mayfield Road
> University of Edinburgh
> Edinburgh EH9 3JR
> UK
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives:
> http://news.gmane.org/gmane.science.biology.informatics.conductor
>
>
>



-- 
Dr. Jose I. de las Heras                      Email: J.delasHeras at ed.ac.uk
The Wellcome Trust Centre for Cell Biology    Phone: +44 (0)131 6513374
Institute for Cell & Molecular Biology        Fax:   +44 (0)131 6507360
Swann Building, Mayfield Road
University of Edinburgh
Edinburgh EH9 3JR
UK