[BioC] RankProd Error Message

[fhong@salk.edu]

> 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.
It depends on which teo conditions you want to compare. Suppose you want
t9o compare MDM vs MDM-INF, then for each of the 4 individuals, you get
ratio of MDM/MDM-INF. YOu can input the resuting 4 ratios to  fucntion RP
with "one-class" option, meaning cl=c(0,0,0,0).
If you want to compare NONINF vs INF regardless of the T condition, I
would suggest you use fucniton RPadvancec to treat (MDM vs MDM-INF) as one
origin and (MDM-T vs MDM-T INF) as the second origin. What you should do
it is get 4 ratios of (MDM/MDM-INF), and 4 ratios of  (MDM-T/MDM-T INF),
pur them togeher as data. Let cl=rep(0,8), origin=c(rep(1,4),rep(2,4))
RPadvance(data, cl, origin)

Let me know if this is not clear.

Bests,
Fangxin



> 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
>>
>> Thanks for any help you can provide!
>>
>> ---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
>
>
>
>


--------------------
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