My perspective on this is that a parsing package should return information in the very basic R data structures (lists, vectors, matrices) and not involve any classes from downstream analysis packages. If we (the parsing package) start to support downstream classes, we will have to depend on a (possible large) number of packages. We (illuminaio) should consider having a convenience function which reads a set of IDAT files into a matrix of intensities; this would be very helpful to many 'users'. The function could check that all IDAT files are of the same type, they all exists, do path expansion etc etc. Then it would be up to a downstream analysis package to take this matrix and wrap their own class structure around it. This should be easy to write (but of course it will take some development time). We would even be able to make the parsing faster by only parsing part of the file (well, that depends on whether it is an encrypted or unencrypted IDAT). Sean, for control probes there is (afaik) no information in the IDAT files; it is just a set of intensities. Best, Kasper On Thu, Nov 28, 2013 at 11:30 AM, Mike Smith <grimbough@gmail.com> wrote: > Hi Gordon, > > Matt has kindly already sent me a copy of his script for creating the EList > object. Combined with Sean's code for generating a LumiBatch and the > script I'm putting together for beadarray I think we have most of the > downstream expression analysis packages covered. My initial hope was that > individual packages would import illuminaio (I like the elegance of having > a single function in the library) and modify their data input routines to > take both the text and IDAT formats. I'll discuss it with Kasper as > illuminaio is ultimately his package. > > Cheers, > > Mike > > > On 28 November 2013 02:20, Gordon K Smyth <smyth@wehi.edu.au> wrote: > > > Dear Sean and Mike, > > > > We (actually Matt Ritchie) have written a nice little function that uses > > readIDAT() to read the IDAT files and assembles the pieces into a data > > object ready for limma analysis. The output is an EList object, but it > > could be an ExpressionSet instead. > > > > It has been suggested that this function could be added to the limma > > package, but I wondered whether the illuminaio package would be the best > > home. > > > > What would you like us to do? > > > > Best wishes > > Gordon > > > > Date: Tue, 26 Nov 2013 09:33:17 -0500 > >> From: Sean Davis <sdavis2@mail.nih.gov> > >> To: "bioconductor@r-project.org" <bioconductor@r-project.org> > >> Subject: [BioC] Illumina Expression analysis from iDAT files > >> > >> Hi, all. > >> > >> I have a set of iDAT files from illumina expression arrays and have been > >> looking around for a workflow that begins with the iDAT files. > Illuminaio > >> can read them and beadarray and lumi packages (to name a couple) can > deal > >> with illumina data normalization and processing, but I didn't find a > >> bridge > >> between the illuminaio output and further processing. I very well may > >> have > >> missed something, but I thought I would check. > >> > >> Thanks, > >> Sean > >> > > > > ______________________________________________________________________ > > The information in this email is confidential and inte...{{dropped:18}} > > _______________________________________________ > Bioconductor mailing list > Bioconductor@r-project.org > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: > http://news.gmane.org/gmane.science.biology.informatics.conductor > [[alternative HTML version deleted]]