Hi Harrys, please post your doubts in the BioC mailing list, so that everyone can benefit of the comments. When you run Agi4x44PreProcess at the end you can get an ExpresionSet object that contains the processed data, using the build.eset function in this way, esetPROC = build.eset(ddPROC, targets, makePLOT = FALSE, annotation.package = "hgug4112a.db) being ddPROC the RGList that contains your processed data, and targets your target file that you should have loaded at the beggining using targets=read.targets(infile="targets.txt"), this is all what you need to start using the limma package. The way I used limma is like this: Imagine that I have a paired design by subject (3 levels) and I want to compare different levels of a Treatment factor with 5 levels. Your target file has to contain this information, for example: FileName Treatment GErep Subject EtOH_P1.txt EtOH 1 1 EtOH_P2.txt EtOH 1 2 EtOH_P3.txt EtOH 1 3 LPS_P1.txt LPS 2 1 LPS_P2.txt LPS 2 2 LPS_P3.txt LPS 2 3 LPS+GW_P1.txt LPS+GW 3 1 LPS+GW_P2.txt LPS+GW 3 2 LPS+GW_P3.txt LPS+GW 3 3 LPS+LG_P1.txt LPS+LG 4 1 LPS+LG_P2.txt LPS+LG 4 2 LPS+LG_P3.txt LPS+LG 4 3 LPS+R_P1.txt LPS+R 5 1 LPS+R_P2.txt LPS+R 5 2 LPS+R_P3.txt LPS+R 5 3 To use limma, first, we need to define the design and contrast matrices. levels.treatment=levels(factor(targets$Treatment)) treatment=factor(as.character(targets$Treatment), levels=levels.treatment) contrasts(treatment)=contr.sum(length(levels.treatment)) levels.subject=levels(factor(targets$Subject)) subject=factor(as.character(targets$Subject), levels=levels.subject) contrasts(subject)=contr.sum(length(levels.subject)) design=model.matrix(~ -1 + treatment + subject ) CM=cbind(L1vsL2=c(-1,1,0,0,0,0,0)) Then, we can fit the linear model and get the contrasts of interest fit=lmFit(esetPROC,design) fit2=contrasts.fit(fit,CM) fit2=eBayes(fit2) You can access the fit2 object to look for genes that are differentially expressed using decideTests(fit2,method="separate",adjust.method="BH",p.value=0.05). This is quite approximately the way I do it. p.- ________________________________ From: cjharrysbackup@gmail.com on behalf of C. J. Harrys Kishore Sent: Wed 2/11/2009 4:05 AM To: Pedro López Romero Subject: Re: Queries on the 4x44PreProcess Package Dear Pedro, Thank you very much for your information. I now would like to do the statistical analyses using limma but I could find no way to load the ProcessedData.txt file for this analysis. It says that I need to load an RG list..Can you please help me with this? Is there any way I can actually load this file for further analysis using limma. Thanks a lot and sorry for all the trouble.. Harrys On Fri, Feb 6, 2009 at 6:37 AM, Pedro López Romero <plopez@cnic.es> wrote: Hi Harrys, there´s not function implemented in the package to save the graphis as ps or pds automatically, although this would be an advance to take into consideration for future realeases. The data that you have is the processed data. Thiese are the data that you have to use to make inferences, either using limma (for differential expression analysis) or some other package depending on what you want to get (cluster analysis, molecular signatures, gene set enrichment analysis). Be aware that the data have not been statistically analyzed in any way, so we don´t have any pvalue so far. to make cluster with the data, i normallly use mapletree (google it), but you might need to use the ctc package (if I recall correctly) to generate an input for mapletree. I wrote the package for the 4x44k data and I have no other sort of Agilent data, althoug I guess It should be possible to use the package with other one-color agi data, as long as the info that is expected to be collected can be found in your data files. best regards p.- ________________________________ From: cjharrysbackup@gmail.com on behalf of C. J. Harrys Kishore Sent: Wed 2/4/2009 11:52 PM To: Pedro López Romero Subject: Queries on the 4x44PreProcess Package Dear Pedro, I have recently started to use the Agi4x44PreProcess package for single color analyses and have the following queries. I am really new to programming as I am a wet lab guy and I really wanted to learn analyses as well. I am new to especially to this package and R so I guess there might be some silly questions as well.: 1. Is there any way to save the graphs and plots for pre and post normalization automatically as JPEG images because the way I am doing now is manually taking screenshots and saving them. 2. The data that I get after normalization in the processeddata.txt file has values what are these? Are they P-values or fold changes or actual expression. I really would like to know as to what the values stand for. 3. Is there another way to visualize the heat maps with say the gene names rather for publication purposes. What I mean is that is there any other tool to visualize this particular data. 4. It is such a great package but I hope you develop one for 2 color analyses as well. 5. How can I use this for a 244K chip analysis? Is it possible in the very first place? I hope that I am not troubling you with these queries and I thank you in advance for your help. Harrys -- C.J.Harrys Kishore PhD. Student Institute of Bioinformatics Discoverer Bldg. 7th Floor International Tech Park Bangalore-560 066 INDIA Ph: +91 80 28416140 Fax: +91 80 28416132 Mob: +91 9880519498 Web: www.ibioinformatics.org -- C.J.Harrys Kishore PhD. Student Institute of Bioinformatics Discoverer Bldg. 7th Floor International Tech Park Bangalore-560 066 INDIA Ph: +91 80 28416140 Fax: +91 80 28416132 Mob: +91 9880519498 Web: www.ibioinformatics.org [[alternative HTML version deleted]]