Hello,

I think the limma guide is pretty good... however, I don't really understand from it how to get very rough data into shape (and what shape is needed) for processing a single channel. For example, I have very simple experiment, I have two replicates of WT and a treatment. I essentially have four columns which correspond to each array. I know which array is which sample, and which row came from where. What I do not know is how to read this into limma and what the corresponding target file should look like.

WT1, WT2, TRMNT1, TRMNT1

#, #, #, #

#, #, #, #

...

For a simple experiment like that, you don't need a targets file, per se. You can just read the data in, and then use the resulting data.frame as input to lmFit.

From the help page for lmFit:

Usage:

     lmFit(object, design=NULL, ndups=1, spacing=1, block=NULL, correlation, weights=NULL,
           method="ls", ...)
     
Arguments:

  object: A matrix-like data object containing log-ratios or
          log-expression values for a series of arrays, with rows
          corresponding to genes and columns to samples.  Any type of
          data object that can be processed by 'getEAWP' is acceptable.

And since this points us to getEAWP, let's look at that:

Usage:

     getEAWP(object)
     
Arguments:

  object: any matrix-like object containing log-expression values.  Can
          be an object of class 'MAList', 'EList', 'marrayNorm',
          'PLMset', 'vsn', or any class inheriting from
          'ExpressionSet', or any object that can be coerced to a
          numeric matrix.

Since a data.frame is an object that can be coerced to a numeric matrix, a data.frame is a suitable input. But do note that R has an unfortunate habit of converting things that it thinks should be factors into factors, if you have any non-numeric things in your input file, you may have to do some coercion of your own to make sure you have numeric values before getEAWP converts to a matrix.

The targets file has one row for each sample, with one column for the sample names and then a column for each experimental variable. The only variable in your experiment is treatment, so you just need a table with 4 rows and 2 columns: sample name and treatment. For your sanity, I recommend that you make sure the row order of the samples in the targets file matches the column order in your data file.