MathGroup Archive 2012

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: Microarray data analysis

  • To: mathgroup at
  • Subject: [mg125094] Re: Microarray data analysis
  • From: Zach Bjornson <bjornson at>
  • Date: Tue, 21 Feb 2012 06:13:24 -0500 (EST)
  • Delivered-to:
  • References: <jh2tae$qf$> <jhstl8$ngf$>

On Feb 19, 11:44 pm, karthikut... at wrote:
> Thanks for your reply. I will explain what I need in detail this time.
> What I have is Agilent single-color microarray analysis datafiles extracted using Agilent feature extraction software (AFE). I ahve TIFF, JPEG and TXT files from AFE. I want to know if mathematics comes with a built-in tool that can accept the above mentioned TIFF image and work on it. If Mathematica can accept the TIFF image, does it also accepts the Grid information file for the TIFF microarray image so that it can find out XV co-ordinates etc.? IF so where can I get the grid info file for my Agilent microarray chip?
> I am not sure, if mathematica can do all these things, what kinds of analysis will it perform over the TIFF file and what will it give as output and in what format? Is it going to be a text file containing probe intensity values? If yes, can I import this output file from Mathematics into GeneSpring?
> I know I am asking too much questions at once. But, pardon me, because these things have been eating my head for a long time and still I could not figure out  a way to analyze the Agilent TIFF file. AFE is of course an options, but I would  like to get more significant results by putting more than one software.
> Thanks a lot.

If you used a fairly recent Agilent scanner in the default mode,
Agilent "TIFF" files are a nonstandard 40 bits. I've yet to find any
software other than AFE that can read these properly. There are two
scenes per color channel, one for the first 32 bits and one for the
remaining 8 bits. The JPEGs are downsampled previews of the images and
are useless for feature extraction.

AFE is highly optimized and performs many steps to arrive at signal
intensities. You'd be hard pressed to come up with a comparable
feature extraction method for Agilent arrays. More importantly, the
work that AFE does is not intrinsic to your final results -- that is,
your data is not likely to be "more significant" if you extract signal
intensities from different software; consider the output from AFE your
"raw data." You're more likely to get different results if you use
different (not necessarily "more significant") downstream software
tools (e.g. GeneSpring vs. others). If you desire "rawer" data, look
at the raw signal intensities columns in the TXT files instead of the
transformed signal intensities.

Hope that helps,

  • Prev by Date: Re: How best to implement a hash table in Mathematica
  • Next by Date: Complex and Solve
  • Previous by thread: Re: Microarray data analysis
  • Next by thread: boundary condition in VariationalBound