[BioC] "AveExp" column has dissappeared from limma topTable
Gordon K Smyth
smyth at wehi.EDU.AU
Tue Apr 3 23:43:23 CEST 2007
Dear Amy,
See
https://www.stat.math.ethz.ch/pipermail/bioconductor/2007-February/016132.html
Note that 'Amean' is the older name for 'AveExpr'.
Best wishes
Gordon
> Date: Mon, 2 Apr 2007 14:23:45 +0100
> From: "Mikhail, Amy" <a.mikhail at abdn.ac.uk>
> Subject: Re: [BioC] "AveExp" column has dissappeared from limma
> topTable
> To: "Marcus Davy" <mdavy at hortresearch.co.nz>
> Cc: bioconductor at stat.math.ethz.ch
>
> Hi Marcus,
>
> "Mosquito" is a subset of my original exprSet (with some probesets taken
> out). So it is just a matrix and now that I look, there is no "Amean"
> element in my fit object. I didn't know that Amean would not be
> derrived for matrices, thanks for pointing that out.
>
> Many thanks,
> Amy
>
> -----Original Message-----
> From: Marcus Davy [mailto:mdavy at hortresearch.co.nz]
> Sent: Monday, April 02, 2007 12:42 AM
> To: Mikhail, Amy; bioconductor at stat.math.ethz.ch
> Subject: Re: [BioC] "AveExp" column has dissappeared from limma topTable
>
>
> Have a look in your 'fit' object and the lmFit code for clues as to what
> is going wrong. When you create a fit object from lmFit, a list element
> called 'Amean' may not have been created. This is what is picked up by
> topTable.
>
> You didn't mention the structure of the argument object 'Mosquito' in
> your lmFit call. Since you have an Affymetrix experiment, is it
> "PLMset", "exprSet" "ExpressionSet" or a matrix (unless you used the
> convert package to make an MAList)?
>
> If 'Mosquito' is a matrix then a list element for Amean is not created
> in your 'fit' object and hence not displayed in topTable
>
>>x <- matrix(1:20, nc=4) #Any old numbers for illustration design <-
>>rep(1,4) lmFit(x, design=rep(1,4))$Amean # Amean not created
> NULL
>> topTable(eBayes(lmFit(x, design)))
> logFC t P.Value adj.P.Val B
> 5 12.5 3.220503 0.001279659 0.006398294 -2.286030
> 4 11.5 2.962863 0.003047925 0.007619813 -2.696287
> 3 10.5 2.705222 0.006825866 0.011376444 -3.072355
> 2 9.5 2.447582 0.014381831 0.017977289 -3.414236
> 1 8.5 2.189942 0.028528445 0.028528445 -3.721928
>
> Whereas an MAList (or "PLMset", "exprSet" "ExpressionSet") should give
> you the AveExpr column. Note: The Amean calculation is not adjusted for
> weights, whereas your coefficients are.
>
>>MA <- new("MAList")
>>MA$M <- x
>>MA$A <- x
>>lmFit(MA, design)$Amean # Amean created
> [1] 8.5 9.5 10.5 11.5 12.5
>> topTable(eBayes(lmFit(MA, design)))
> logFC AveExpr t P.Value adj.P.Val B
> 5 12.5 12.5 3.220503 0.001279659 0.006398294 -2.286030
> 4 11.5 11.5 2.962863 0.003047925 0.007619813 -2.696287
> 3 10.5 10.5 2.705222 0.006825866 0.011376444 -3.072355
> 2 9.5 9.5 2.447582 0.014381831 0.017977289 -3.414236
> 1 8.5 8.5 2.189942 0.028528445 0.028528445 -3.721928
>
> Hope that helps.
>
>
> Marcus
>
> On 31/3/07 9:57 AM, "Mikhail, Amy" <a.mikhail at abdn.ac.uk> wrote:
>
>> Dear Bioconductors,
>>
>> I have a problem with the "AveExp" column in topTable. Usually it
>> appears in my toptable output, but during a limma session recently I
>> created one toptable (which did have the "AveExp" column in it) and
>> then another toptable (which was a subset). In the second topTable,
>> the AveExp column was missing. I have tried reloading R, and
>> reinstalling limma but the column seems now to be permanently missing
>> from my topTables.
>>
>> There was no error message or warning, so I really don't know what
>> I've done ;-(. Sample code is below:
>>
>>> targets <- readTargets("BaseAgeAffyTargets.txt")
>>> targets
>> Name FileName Target
>> 1 R2s2 AM_041006_R2S2v1_Plasmodium_Anopheles.CEL Adults age 15 days
>> 2 R2s6 AM_041006_R2S2_Plasmodium_Anopheles.CEL Adults age 43 days
>> 3 R3s2 AM_041006_R3S2_Plasmodium_Anopheles.CEL Adults age 15 days
>> 4 R3s6 AM_041006_R3S6_Plasmodium_Anopheles.CEL Adults age 43 days
>> 5 R4s2 AM_041006_BAR4Sii_Plasmodium_Anopheles.CEL Adults age 15 days
>> 6 R4s6 AM_041006_R4S6_Plasmodium_Anopheles.CEL Adults age 43 days
>>>
>>> design <- cbind(M15d=c(1,0,1,0,1,0),M43d=c(0,1,0,1,0,1))
>>> fit <- lmFit(Mosquito, design)
>>> cont.matrix <- makeContrasts(M15dvsM43d=M43d-M15d, levels=design)
>>> fit2 <- contrasts.fit(fit, cont.matrix)
>>> fit2 <- eBayes(fit2)
>>> Lowvar.toptable <- topTable(fit2[ Lowvar.genes, ], number=100,
>> adjust="BH")
>>
>>> Lowvar.toptable[1:8,]
>> ID logFC t
>> P.Value adj.P.Val B
>> 8 Ag.2L.2099.0_CDS_at 0.0525816225 0.613337707
>> 0.5608673 0.9967802 -5.844476
>> 57 Ag.UNKN.1771.0_CDS_at 0.0514293902 0.602622882
>> 0.5675464 0.9967802 -5.851185
>> 46 Ag.3R.2483.0_CDS_at 0.0460176987 0.540864707
>> 0.6069533 0.9967802 -5.887786
>> 24 Ag.2R.3361.0_CDS_at -0.0461903587 -0.540335772
>> 0.6072973 0.9967802 -5.888084
>> 60 Ag.UNKN.363.0_CDS_at -0.0394899270 -0.464386924 0.6577950
>> 0.9967802 -5.928101
>> 64 Ag.X.217.0_CDS_a_at -0.0393985675 -0.455020526
>> 0.6641680 0.9967802 -5.932647
>> 58 Ag.UNKN.1863.0_CDS_at 0.0359376810 0.415978011
>> 0.6910563 0.9967802 -5.950663
>> 30 Ag.3L.1321.0_CDS_at -0.0334439607 -0.387588909
>> 0.7109213 0.9967802 -5.962804
>>
>> I also note that "A" is also missing from my fit object;
>>
>>> fit2$Amean
>> NULL
>>
>> I would be most grateful for any suggestions as to what could be
>> happening here....
>> I realise that I am using R2.4.0 with a more recent version of limma,
>> but this has not caused me any problems up until now.
>>
>> Looking forward to any suggestions,
>> Best wishes,
>> Amy
>>
>>
>>> sessionInfo()
>> R version 2.4.0 (2006-10-03)
>> 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" "methods" "stats" "graphics"
>> "grDevices" "utils" "datasets" "base"
>>
>> other attached packages:
>> mixtools mvtnorm agahomology affyPLM gcrma
> matchprobes
>> affydata annaffy KEGG GO genefilter
>> "0.1.0" "0.7-5" "1.14.2" "1.10.0" "2.6.0"
> "1.6.0"
>> "1.10.0" "1.6.1" "1.14.0" "1.14.0" "1.12.0"
>> survival limma geneplotter annotate affy
> affyio
>> Biobase
>> "2.30" "2.9.8" "1.12.0" "1.12.1" "1.12.2"
> "1.2.0"
>> "1.12.2"
>>
>> **********************************************************************
>> **
>> *********************************************************
>>
>> Amy Mikhail
>> Research student
>> University of Aberdeen
>> Zoology Building
>> Tillydrone Avenue
>> Aberdeen
>> AB24 2TZ
>> Scotland
>> Email: a.mikhail at abdn.ac.uk
>> Phone: 00-44-1224-273256 (office)
----------------------------
Dr Gordon K Smyth, Research Fellow, Bioinformatics,
Walter and Eliza Hall Institute of Medical Research,
1G Royal Parade, Parkville, Vic 3050, Australia
Tel: (03) 9345 2326, Fax (03) 9347 0852,
Email: smyth at wehi.edu.au, www: http://www.statsci.org
More information about the Bioconductor
mailing list