Hi Bob, The lumiMouseAll.db (or lumiHumanAll.db and lumiRatAll.db) only includes the annotation information of the probe. It does not include the Illumina ID and nuID mapping information. The reason is that the recent versions of BeadStudio always output the probe sequence information together with the data. So the lumiR can automatically convert the probe sequences as nuIDs without needing any libraries. For your case, because your data is in old version (v1.1), it may not include the probe sequence in the file. So you need to map the Illumina ID to nuID first by using package lumiMouseV1 like this: x.lumi <- lumiR(fileName,lib='lumiMouseV1') Later on, if you want to retrieve the annotation information, you can directly use lumiMouseAll.db, which is in the same format as other Bioconductor annotation .db packages. We are also trying to set up an online service to mapping different Illumina IDs to nuIDs and then retrieve annotation information. Tell me if you have other questions. Pan On 5/2/08 12:35 PM, "Robert Searles" <searlesr at="" ohsu.edu=""> wrote: > Pan, > > I'm using R 2.70 and BioConductor 2.2. > > I previously had used lumiMouseV1 as the annotations file, but for this > set-up, I used the new lumiMouseAll.db library. The data file is for a > set of MouseWG6 v1.1 arrays. > > I run the lumi protocol as follows: > >> library(lumi) >> library(lumiMouseAll.db) > Loading required package: AnnotationDbi > Loading required package: DBI > Loading required package: RSQLite >> fileName <- "c:/RH248_1_Group_Probe_Profile.txt" >> x.lumi <- lumiR(fileName,lib='lumiMouseAll.db') > Warning messages: > 1: In addNuId2lumi(x.lumi, lib = lib) : > lumiMouseAll.db does not include nuID conversion information! > 2: In addNuId2lumi(x.lumi, lib = lib) : > Please provide the annotation file or lumi annotation library! > > I get the two warning messages, which indicate that there is another > file needed to get the nuID conversions. Is that correct or am I doing > something wrong? > > Bob > > > >>>> "Pan Du" <dupan at="" northwestern.edu=""> 5/1/2008 9:12:04 AM >>> > Hi Bob, > > Yes. It is the lumiMouseAll.db package. The package is in the > Bioconductor > developing version (Bioc 2.2), which will be released very soon. You > need to > use R2.7 if you use biocLite to install it. You can also manually > download > and install it if you are using early R versions. Tell me if you meet > any > problems when using it. We are also setting up a website for Illumina > ID > mapping. > > > Pan > > > On 5/1/08 11:03 AM, "Robert Searles" <searlesr at="" ohsu.edu=""> wrote: > ------------------------------------------------------ Pan Du, PhD Research Assistant Professor Biomedical Informatics Center Northwestern University 750 N. Lake Shore Drive, 11-176 Chicago, IL 60611 Office (312) 503-2360 dupan at northwestern.edu

Hi Bob, I guess the chip you use is Version 1.1 R2, which includes some new probes not included in the old releases (lumiMouseV1 was based on version 1.1 R0 and version 1.0). I am not sure how to output annotation information in the BeadStudio in our microarray core here. I will check with them later. Another way is to unzip the manifest file (.bgx), which basically is zipped by gzip. You can open the unzipped file (basically it is a text file) and check whether there is a probe sequence column. If so, you can use this file to convert the Ilumina IDs as nuIDs by using: x.lumi = lumiR(filename, convertNuID=FALSE) x.lumi = addNuId2lumi(x.lumi, annotationFile='MouseWG-6_V1_1_R2_11234304_A', annotationColName=c(sequence='Probe_Sequence', probe='Probe_Id')) Here suppose 'MouseWG-6_V1_1_R2_11234304_A' is the manifest file you use. Tell me if this does not work for you. Have a nice weekend, Pan On 5/2/08 2:56 PM, "Robert Searles" <searlesr at="" ohsu.edu=""> wrote: > Pan, > > Thanks for the quick reply and I apologize for being a nudge about > this. > > I'm using the most current version of BeadStudio and of the BeadStudio > Gene Expression Module (I just verified this to be correct). The data > for the arrays is derived from the new .bgx files, not the old .csv > files, though I don't think that has much relevance here. BeadStudio > v3.1.3 is not, by default, exporting the sequence data along with the > rest of the data; at least my installation is not. > > Given that I'm using the newest software and manifests, I guess a > better question for me to ask is "how do you output the correct data > from BeadStudio?" I've been using either the GeneSpring GX Group Probe > Profile option or Sample Probe Profile on the File menu. Is this the > correct method or is this where I'm making my first mistake? The data > that I'm getting admittedly doesn't look like the example in the > documentation - the header format is a bit different and the block > designators don't come through at all. > > > When I use lumiMouseV1 with the Mousev1.1 arrays, only the probeids > that are also on the older V1 arrays are converted to nuIDs. The new > probes that occur only on the v1.1 array (about 2096 probes) instead > retain their probeids. So, when I convert the lumi.N file to an exprs > object and output it using write.table, the first column is a mix of > nuIDs (for 95% of the array) and probeids (the remaining 5% of the > array). On the off chance that my copy of lumiMouseV1 was outdated, I > reloaded it and then also set up BioConductor on a computer that had > never seen R before and did a clean install. In all cases, I was unable > to convert all probeids to nuIDs using lumiMouseV1. My hope was that > the new lumiMouseAll.db library would solve this problem. > > Bob > > > >