[BioC] can't use weights in single channel analysis?

[Jenny Drnevich]

Hello,

I have some 2-color microarrays that I need to analyze as single 
channel because some of the conditions I want to compare are 
unconnected. I've been trying to follow the example in the 
limmaUsersGuide() on "Separate Channel Analysis of Two-Color Data". 
However, it appears that neither 
normalizeBetweenArrays(method="Aquantile") nor the lmscFit() function 
use the weights in the MAList object. Four of my 19 arrays had lower 
hybridization efficiency on the bottom quarter of the array (courtesy 
of an air bubble), leading to very low R and G values for these 
spots. The rest of the values on the arrays seem fine, so I don't 
want to completely throw these arrays out, just the bad spots which 
I've given weights==0.

Within normalizeBetweenArrays(), the internal function 
normalizeQuantiles() can handle missing values, so I was able to do 
an appropriate Aquantile normalization by replacing the A values for 
spots with weight==0 with NA. However, you can't have missing values 
when calculating intraspotCorrelation(), nor can you have missing 
values for lmscFit(). In looking through the code of lmscFit(), it 
doesn't use weights, even if they are in the MAList object! So as is, 
lmscFit() won't give me the proper coefficients for the spots that 
have weight==0 on one or more arrays. I'm not sure I'm able to modify 
the code of lmscFit() to use the weights, or even if it can be 
modified to use the weights.

I've though about manually creating a matrix of my R & G values, 
using NA for spots with weight==0, but I'm not sure how to model the 
correlation between the R & G channels on an array. Anybody have any 
suggestions on how I can appropriately analyze my data?

Thanks,
Jenny

  sessionInfo()
R version 2.10.1 (2009-12-14)
i386-pc-mingw32

locale:
[1] LC_COLLATE=English_United States.1252  LC_CTYPE=English_United 
States.1252
[3] LC_MONETARY=English_United States.1252 
LC_NUMERIC=C
[5] LC_TIME=English_United States.1252

attached base packages:
[1] grid      stats     graphics  grDevices 
datasets  utils     methods   base

other attached packages:
  [1] 
statmod_1.4.1        limma_3.2.1          affyQCReport_1.24.0 
lattice_0.17-26      xtable_1.5-6
  [6] 
simpleaffy_2.22.0    genefilter_1.28.2    made4_1.20.0 
scatterplot3d_0.3-29 gplots_2.7.4
[11] 
caTools_1.10         bitops_1.0-4.1       gdata_2.6.1 
gtools_2.6.1         RColorBrewer_1.0-2
[16] ade4_1.4-14          affyPLM_1.22.0       preprocessCore_1.8.0 
gcrma_2.18.1         affycoretools_1.18.0
[21] 
KEGG.db_2.3.5        GO.db_2.3.5          RSQLite_0.8-0 
DBI_0.2-5            AnnotationDbi_1.8.1
[26] affy_1.24.2          Biobase_2.6.1        RWinEdt_1.8-2

loaded via a namespace (and not attached):
  [1] 
affyio_1.14.0      annaffy_1.18.0     annotate_1.24.0 
biomaRt_2.2.0      Biostrings_2.14.10
  [6] 
Category_2.12.0    GOstats_2.12.0     graph_1.24.1 
GSEABase_1.8.0     IRanges_1.4.9
[11] 
RBGL_1.22.0        RCurl_1.2-1        splines_2.10.1 
survival_2.35-7    tools_2.10.1
[16] XML_2.6-0


Jenny Drnevich, Ph.D.

Functional Genomics Bioinformatics Specialist
W.M. Keck Center for Comparative and Functional Genomics
Roy J. Carver Biotechnology Center
University of Illinois, Urbana-Champaign

330 ERML
1201 W. Gregory Dr.
Urbana, IL 61801
USA

ph: 217-244-7355
fax: 217-265-5066
e-mail: drnevich at illinois.edu