[BioC] siggenes and permutations

[Louisa A Rispoli/AS/EXP/UTIA]

Hi all-

We are investigating SAM and EBAM analysis for our data and run into a
problem that reading the archives has not helped any. We performed SAM on a
data set and noticed the permutations randomly selected 100 to generate the
results. We were curious what that meant and if had anything similar to
iterations in SAS (more usually is better). So we tried more and discovered
that it was not better necessarly to us (increased FDR values for same
deltas). And since we do not exactly understand what permutations exactly
are doing in this case we wanted to check with the experts on their opinion
if it should be changed are left at the default of 100.

Thanks for the time and energy

 Louisa



"If we knew what we were doing, it wouldn't be called Research." - Albert
Einstein

Louisa Rispoli, Ph.D. Reproductive Physiology
Department of Animal Science
University of Tennessee, Knoxville
A105 Johnson Animal Research and Teaching Unit
1750 Alcoa Highway
Knoxville, TN 37920
phone:(865) 946-1874
fax:(865) 946-1010
email: lrispoli at utk.edu

> library(siggenes)
Loading required package: multtest
Loading required package: survival
Loading required package: splines
> trt <- pData(Poly.rma.filter)$Treatment
> trt
 [1] "Ctrl" "Ctrl" "Ctrl" "Ctrl" "Ctrl" "Ctrl" "Ctrl" "Ctrl" "HS"   "HS"
"HS"   "HS"   "HS"   "HS"   "HS"   "HS"
> clPoly <- ifelse(trt=="Ctrl", 0, 1)
> clPoly
 [1] 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1
> dataPoly.rma <- exprs(Poly.rma.filter)

> sam.out <-sam(dataPoly.rma, clPoly, rand=123456)

We're doing 12870 complete permutations
and randomly select 100 of them.

> sam.out
SAM Analysis for the Two-Class Unpaired Case Assuming Unequal Variances

   Delta    p0   False Called   FDR
1    0.1 0.666 7190.38   8620 0.556
2    0.2 0.666 6265.49   7932 0.526
3    0.3 0.666 2303.43   3891 0.394
4    0.4 0.666  299.36    660 0.302
5    0.5 0.666   90.83    246 0.246
6    0.6 0.666   34.53    117 0.197
7    0.7 0.666   10.92     44 0.165
8    0.8 0.666    1.06      6 0.118
9    0.9 0.666       0      0     0
10   1.0 0.666       0      0     0

> sam.out1 <- sam(dataPoly.rma, clPoly, rand=123456, B=1000)

We're doing 12870 complete permutations
and randomly select 1000 of them.

> sam.out1
SAM Analysis for the Two-Class Unpaired Case Assuming Unequal Variances

   Delta    p0    False Called   FDR
1    0.1 0.675 7371.715   8705 0.571
2    0.2 0.675 5928.818   7584 0.527
3    0.3 0.675 1275.681   2233 0.385
4    0.4 0.675  255.636    549 0.314
5    0.5 0.675   87.618    228 0.259
6    0.6 0.675   37.511    116 0.218
7    0.7 0.675   12.686     44 0.195
8    0.8 0.675    2.287      8 0.193
9    0.9 0.675        0      0     0
10   1.0 0.675        0      0     0
> sessionInfo()
R version 2.7.1 (2008-06-23)
i386-pc-mingw32

locale:
LC_COLLATE=English_United States.1252;LC_CTYPE=English_United
States.1252;LC_MONETARY=English_United
States.1252;LC_NUMERIC=C;LC_TIME=English_United States.1252

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

other attached packages:
 [1] siggenes_1.14.0      multtest_1.20.0      affycoretools_1.12.0
annaffy_1.12.1       KEGG.db_2.2.0        biomaRt_1.14.0
 [7] RCurl_0.9-3          GOstats_2.6.0        Category_2.6.0
RBGL_1.16.0          annotate_1.18.0      xtable_1.5-2
[13] GO.db_2.2.0          AnnotationDbi_1.2.2  RSQLite_0.6-9
DBI_0.2-4            graph_1.18.1         limma_2.14.5
[19] bovinecdf_2.2.0      simpleaffy_2.16.0    gcrma_2.12.1
matchprobes_1.12.0   genefilter_1.20.0    survival_2.34-1
[25] affy_1.18.2          preprocessCore_1.2.0 affyio_1.8.0
Biobase_2.0.1

loaded via a namespace (and not attached):
[1] cluster_1.11.11 XML_1.95-3