[BioC] problems with readIllumina() for many BeadChips

[Ina Hoeschele]

thank you very much, Mike. 
The . csv files contain Illumina code, N (number of beads/bead type I assume), mean GRN, and dev GRN.
Ina



----- Original Message -----
From: "Mike Smith" <grimbough at gmail.com>
To: "Ina Hoeschele" <inah at vbi.vt.edu>
Cc: bioconductor at stat.math.ethz.ch
Sent: Thursday, February 17, 2011 9:26:07 AM
Subject: Re: [BioC] problems with readIllumina() for many BeadChips

Hi Ina,

I've removed the offending line of code from the createTargetsFile function,
which was causing the errors you've been seeing.  Version 2.0.4 should work
without requiring the section names to be specified. The directory scanning
functionality is something I've been working on a lot in the past week and
the development version of the package should be more resilient to having a
variety of files in a directory.

Out of interest, what data is stored in the .csv files you have? I wasn't
aware that BeadScan produced anything with that extension, so if there's
useful data in there I may need to make some more allowances in the
directory scanning function.

As far as reading multiple chips goes, we tend to avoid it as you start to
run out of memory very quickly once you call some of the processing
functions.  Beadarray also only performs outlier removal, intensity
normalization etc. within a chip, rather than between them, so there is
little advantage in reading everything in in one go. Our approach is
generally to list all the directories in a vector and loop through it, or to
farm each directory out on a cluster. We then combine them once they've been
summarized.

The additional arguments to readIllumina mentioned in the help file are
passed to functions such as readBeadLevelTextFile allowing you to specify
seperators etc. I agree the documentation is lacking (non existant), and
we'll try to make it clearer shortly.

I hope this is of some help,
On 15 Feb 2011 22:06, "Ina Hoeschele" <inah at vbi.vt.edu> wrote: