[BioC] toptable with multiple coefs [was Time course experiment....]

[Gordon Smyth]

I give results to many of my collaborators in exactly the same way 
(except for sorting by gene ID, see below). They like to have all the 
results in an Excel spreadsheet so that they can play around with it 
a bit themselves. I use

   write.fit(fit)

to create such a file. Does this do what you want?

Regarding the gene IDs, there isn't a way for limma functions to sort 
results by gene ID, because there isn't a requirement that the 
'genes' component of a MArrayLM object contains a column which is a 
unique probe identifier. Even if such a column does exist, there is 
no requirement that it has to be called "ID". So limma doesn't have a 
way to know which column to use to sort by gene ID.

However with any specific data set, you will know which column is 
your ID column. So you could use

  o <- order( fit$genes$ID )
  write.fit( fit[o,] )

Cheers
Gordon

At 11:38 AM 4/01/2006, Naomi Altman wrote:
>Just as this discussion got underway, I met with my current 
>microarray collaborator.
>
>She asked that
>
>a) The results be sorted by gene id
>b) all the coefficients and their p-values be printed out to the 
>same (big) table
>c) all the gene annotations be added to the table
>
>Very timely.
>
>--Naomi