[R] Is it normal that normalize.loess does not tolerate a single NA value?

[Emmanuel Levy]

Dear all,

I have been using normalize.loess and I get the following error
message when my matrix contains NA values:

> my.mat = matrix(nrow=100, ncol=4, runif(400) )
> my.mat[1,1]=NA
> my.mat.n = normalize.loess(my.mat, verbose=TRUE)
Done with 1 vs 2 in iteration 1
Done with 1 vs 3 in iteration 1
Done with 1 vs 4 in iteration 1
Done with 2 vs 3 in iteration 1
Done with 2 vs 4 in iteration 1
Done with 3 vs 4 in iteration 1
1 0.317319
Warning messages:
1: In means[, j] + aux :
  longer object length is not a multiple of shorter object length
2: In means[, k] - aux :
  longer object length is not a multiple of shorter object length
...

Is that normal? If not, do you have any suggestion to avoid it? I'm
scared that this introduces abnormalities in the normalization.

Thanks for your help,

Emmanuel


> sessionInfo()
R version 2.8.0 (2008-10-20)
x86_64-pc-linux-gnu

locale:
LC_CTYPE=en_CA.UTF-8;LC_NUMERIC=C;LC_TIME=en_CA.UTF-8;LC_COLLATE=en_CA.UTF-8;LC_MONETARY=C;LC_MESSAGES=en_CA.UTF-8;LC_PAPER=en_CA.UTF-8;LC_NAME=C;LC_ADDRESS=C;LC_TELEPHONE=C;LC_MEASUREMENT=en_CA.UTF-8;LC_IDENTIFICATION=C

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

other attached packages:
 [1] limma_2.16.3        GOstats_2.4.0       Category_2.4.0
 [4] genefilter_1.22.0   survival_2.34-1     RBGL_1.16.0
 [7] annotate_1.20.1     xtable_1.5-4        GO.db_2.2.0
[10] AnnotationDbi_1.2.2 RSQLite_0.6-9       DBI_0.2-4
[13] graph_1.18.0        YEAST_2.0.1         affy_1.20.0
[16] Biobase_2.2.1

loaded via a namespace (and not attached):
[1] affyio_1.10.1        cluster_1.11.11      preprocessCore_1.4.0
[4] tcltk_2.8.0