[BioC] can't use weights in single channel analysis?
drnevich at illinois.edu
Wed Mar 10 22:28:04 CET 2010
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?
R version 2.10.1 (2009-12-14)
 LC_COLLATE=English_United States.1252 LC_CTYPE=English_United
 LC_MONETARY=English_United States.1252
 LC_TIME=English_United States.1252
attached base packages:
 grid stats graphics grDevices
datasets utils methods base
other attached packages:
statmod_1.4.1 limma_3.2.1 affyQCReport_1.24.0
simpleaffy_2.22.0 genefilter_1.28.2 made4_1.20.0
caTools_1.10 bitops_1.0-4.1 gdata_2.6.1
 ade4_1.4-14 affyPLM_1.22.0 preprocessCore_1.8.0
KEGG.db_2.3.5 GO.db_2.3.5 RSQLite_0.8-0
 affy_1.24.2 Biobase_2.6.1 RWinEdt_1.8-2
loaded via a namespace (and not attached):
affyio_1.14.0 annaffy_1.18.0 annotate_1.24.0
Category_2.12.0 GOstats_2.12.0 graph_1.24.1
RBGL_1.22.0 RCurl_1.2-1 splines_2.10.1
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
1201 W. Gregory Dr.
Urbana, IL 61801
e-mail: drnevich at illinois.edu
More information about the Bioconductor