[BioC] again: warning message for duplicateCorrelation (Limma)

[dorthe.belgardt at medisin.uio.no]

Hi again!

I thought it might be better to add some information to my message
concerning my problem using the duplicateCorrelation function in Limma
(see below).

The arrays are used are printed in duplicates with a spacing of 15000 (so
in total 30000 features per array) and I did the imageprocessing using
GenePixPro6.1.

This is what I have done before applying the dupcor-function:

> targets=readTargets("Targets_basicSat.txt")
> targets
   SlideNumber          FileName Cy3 Cy5
1            1 3096_basicSat.gpr ref   A
2            2 3079_basicSat.gpr   A ref
3            3 3089_basicSat.gpr ref   A
4            4 3081_basicSat.gpr   A ref
5            5 3071_basicSat.gpr ref   B
6            6 3082_basicSat.gpr   B ref
7            7 3085_basicSat.gpr ref   B
8            8 8268_basicSat.gpr   B ref
9            9 7829_basicSat.gpr ref   C
10          10 3086_basicSat.gpr   C ref
11          11 7823_basicSat.gpr ref   C
12          12 7826_basicSat.gpr   C ref
13          13 3090_basicSat.gpr ref   D
14          14 3091_basicSat.gpr   D ref
15          15 3092_basicSat.gpr ref   D
16          16 7827_basicSat.gpr   D ref

Every other slide is a dyeswapped technical replicate and per "group"
(A,B,C,D) there are 2 biological replicates.

> K=read.maimages(targets$FileName, source="genepix.median",
wt.fun=wtflags(0))
> types=readSpotTypes("SpottypesGAPDH.txt")
> Status=controlStatus(types, K)
> K$genes$Status=Status
> K2=backgroundCorrect(K, method=”subtract”)
> K2=normalizeWithinArrays(K2, method="median")
> design=modelMatrix(targets, ref="ref")
> design
      A  B  C  D
 [1,]  1  0  0  0
 [2,] -1  0  0  0
 [3,]  1  0  0  0
 [4,] -1  0  0  0
 [5,]  0  1  0  0
 [6,]  0 -1  0  0
 [7,]  0  1  0  0
 [8,]  0 -1  0  0
 [9,]  0  0  1  0
[10,]  0  0 -1  0
[11,]  0  0  1  0
[12,]  0  0 -1  0
[13,]  0  0  0  1
[14,]  0  0  0 -1
[15,]  0  0  0  1
[16,]  0  0  0 -1

>cor=duplicateCorrelation(K2, design=design, ndups=2, spacing=15000,
weights=K2$weights)

And at this point, I get the warning message "NaNs produced in:
sqrt(dfitted.values)" and I do not know if I should simply procede or if
this warning indicates a major problem.

I hope somebody can help and that this is sufficient information.
Kind regards,
Dorthe


Hi,

when using the dulicateCorrelation function I get the following warning
message: "NaNs produced in: sqrt(dfitted.values)"
I have used the function before and usually got the "too much damping"
warning message, which I know can be ignored. But now I get the warning
message metioned above and don't know whether to ignore it or not.

Thanks for any help in advance,
Dorthe

-- 
Dorthe Belgardt
Institute of Basic Medical Sciences
Department of Physiology
P.O. Box 1103 Blindern
0317 Oslo
Norway







-- 
Dorthe Belgardt
Institute of Basic Medical Sciences
Department of Physiology
P.O. Box 1103 Blindern
0317 Oslo
Norway
Tel: 0047-22851295