Dahlquist:Notebook/Microarray Data Analysis/2008/10/07

From OpenWetWare

< Dahlquist:Notebook | Microarray Data Analysis | 2008 | 10
Revision as of 20:43, 13 October 2008 by Kam D. Dahlquist (Talk | contribs)
(diff) ←Older revision | Current revision (diff) | Newer revision→ (diff)
Jump to: navigation, search

Image:HeaderBlueBig.gif Home        Research        Protocols        Notebook        People        Publications        Courses        Contact       


Microarray Data Analysis Main project page
Previous entry      Next entry

Kevin's Edge Analysis for 10/7/08

  • Previously, I successfully compared the wt and Cin5 datasets (10/2/08). Now I am attempting to determine how to analyze the individual datasets.
  • The EDGE manual states that "Class Variable should be used to select the variable that identifies the biological groups to be compared. If only one biological group has been sampled select None."
    • Thus the only way to perform within class differential expression appears to be to select None
  • Then the covariate giving time points must be chosen as timepoint
  • Choosing covariate corresponding to individuals:
    • The manual states that "If individuals were sampled at more than one time point (making these arrays dependent), then the variable denoting which individual corresponds to which array should be selected).
    • In the previous analysis, Flask was chosen. Doing this now, however, would treat the entire dataset as one and would make no distinction between wt and cin5, I believe.
    • The second option is to select strain here, but I'm not sure what this would compare. If it did treat the wt and Cin5 separately, however, I don't think it would distinguish between Flasks.
  • So there appear to be a couple ways to fix this problem:
    • Maybe I am missing something, and we could attempt to analyze the data using either the Flask or Strain covariate as the covariate corresponding to individuals and see what happes.
    1. I attempted to look online on the Google Edge discussion page and found someone with the exact same question, but no one had answered it. I posted a response of my own and will see if anyone responds.
    2. By creating new gene and covariate files with only the Cin5 or wt data, we could ensure that only the individual dataset is being compared.

--Kevin C. Entzminger 17:03, 7 October 2008 (EDT)

Comments

  • I looked back in my notes and saw that I had made new gene and covariate files for wt-only and dCIN5-only (like you proposed in #2 above). These probably got transferred over to your login, but let's double check that. Kam D. Dahlquist 20:43, 13 October 2008 (EDT)



Personal tools