# [BioC] RankProd Error Message

Ettinger, Nicholas nicholas-ettinger at uiowa.edu
Wed Jan 4 20:41:11 CET 2006

```Thank you for the information!  I think that I almost have it.  I didn't
realize that I needed to input a matrix-format data set.

With regards the 'two-class' vs. pairwise question, I'm still not quite
sure which I need.  My statistics background is not so solid and so I
don't have good insight here.

My samples below correspond (i.e. 1 -> 4) to 4 different individuals
sampled on different days.  In other words, each individual experiment
consisted of cells from one individual tested under 4 conditions (MDM
only, MDM-INF, MDM-T only, MDM-T INF), and this whole experiment was
repeated (on different days) with 4 people overall.  I.e. 4 people x 4
conditions each time @ 1 array per condition = 16 arrays.

With that extra information, any insights on how to set up the 'cl'
matrix?

Or should I really be taking an average of each of the 4 matrices for
each condition first?  So that I end up with 4 "average" data sets?

Thanks!  I have been really struggling with this concept.

---Nick

-----Original Message-----
From: fhong at salk.edu [mailto:fhong at salk.edu]
Sent: Wednesday, January 04, 2006 1:10 PM
To: Ettinger, Nicholas
Cc: bioconductor at stat.math.ethz.ch
Subject: Re: [BioC] RankProd Error Message

Since there are two classed in your data, the class label(eset.cl)
should
have two categories only. Try to use
eset.cl <- rep(c(rep(0,4),rep(1,4)),2)
> eset.cl
[1] 0 0 0 0 1 1 1 1 0 0 0 0 1 1 1 1

then your class "INF" would be class 1 and "NONINF" would be class 2.
And
make sure your eset_rma is not a affybatch object, but a matrix-format
data set.

But if your experiment is pairwise design,meaning 1 and -1 is a pair, 2
and -2 is a pair and so on, then you need to get ratio of 1/-1, 2/-2,...
and use the "one-class" option in function RP.

Let me know what you get.

Bests,
Fangxin

> Hello all!
>
> I'm trying to analyze 16 Affy arrays with SAM, Limma and RankProd to
> look at differential expression.
>
> I got the following error message from the RP function.  Any help as
to
> what it means?
>
>> RP.out<- RP(eset_rma, eset.cl, num.perm = 100, logged=TRUE,
> na.rm=FALSE, plot=FALSE, gene.names=eset.gnames)
> Error in if (total.sam != total.sam2) stop("Number of classes should
> match the columns in the data") :
>         argument is of length zero
>
> I have included below the phenotype data of eset_rma and the listing
of
> eset.cl:
>
>> pData(eset_rma)
>                    SAMPLE_NUM TREATMENT TCELLS
> NE_MDM-INF1.CEL             1       INF      N
> NE_MDM-INF2.CEL             2       INF      N
> NE_MDM-INF3.CEL             3       INF      N
> NE_MDM-INF4.CEL             4       INF      N
> NE_MDM-T-1.CEL              5    NONINF      Y
> NE_MDM-T-2.CEL              6    NONINF      Y
> NE_MDM-T-3.CEL              7    NONINF      Y
> NE_MDM-T-4.CEL              8    NONINF      Y
> NE_MDM-T-INF-1.CEL          9       INF      Y
> NE_MDM-T-INF-2.CEL         10       INF      Y
> NE_MDM-T-INF-3.CEL         11       INF      Y
> NE_MDM-T-INF-4.CEL         12       INF      Y
> NE_MDM1.CEL                13    NONINF      N
> NE_MDM2.CEL                14    NONINF      N
> NE_MDM3.CEL                15    NONINF      N
> NE_MDM4.CEL                16    NONINF      N
>
>> eset.cl
>  [1]  1  2  3  4 -5 -6 -7 -8  5  6  7  8 -1 -2 -3 -4
>
>
> ---Nick Ettinger
>
> 	[[alternative HTML version deleted]]
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
>
>

--------------------
Fangxin Hong  Ph.D.
Plant Biology Laboratory
The Salk Institute
10010 N. Torrey Pines Rd.
La Jolla, CA 92037
E-mail: fhong at salk.edu
(Phone): 858-453-4100 ext 1105

```