HI,

I am new to R.I am using a windows base system. I am following the tut by Daniel Swan. After writing the following command

> gene.symbols <- getSYMBOL(probeset.list$ID, "hgu133plus2")

get this eror:

Error in unlist(lookUp(x, data, "SYMBOL")) :
  error in evaluating the argument 'x' in selecting a method for function
'unlist': Error in lookUp(x, data, "SYMBOL") : No keys provided

Can any one help.
rabia

Hi Rabia,
I'm new on R like you but i have some experience with this type of analysis
So in your case verify that your probeset.list contain some gene DE
!!
att the first tap this :

head(probeset.list)  and told me the result !!

Dear all,

I came up to the same error, following the same tutorial.

As suggested above, I tried "head(probeset.list)  ".

Here are the results:

            logFC  AveExpr         t      P.Value    adj.P.Val        B
204779_s_at  7.367790 4.171707  72.77347 3.284937e-15 8.969850e-11 20.25762
207016_s_at  6.936667 4.027733  57.39252 3.694641e-14 5.044293e-10 19.44987
209631_s_at  5.192949 4.003992  51.24892 1.170273e-13 1.065182e-09 18.96660
242809_at    6.433238 4.168870  48.51842 2.043082e-13 1.394710e-09 18.70852
205893_at    4.480331 3.543714  40.56477 1.261400e-12 6.888757e-09 17.75050
204882_at   -5.353547 6.513690 -34.70646 6.141288e-12 2.395629e-08 16.77396

Any idea ?

Thanks,

Alexandre

[EDIT]

I made it work thanks to a comment from a thread in another forum.

http://stackoverflow.com/questions/22970546/microarray-limma-package-in-toptable-function-dont-assign-id-for-probsets-colu

When "probeset.list$ID"  is changed to "rownames(probeset.list)", the error is fixed. This error was mmost likely to an update of limma, as far as I understood

This question apparently relates to Daniel Swan's 2011 tutorial about using the simpleaffy and limma packages to analyse Affymetrix microarray data:

  http://bioinformatics.knowledgeblog.org/2011/06/20/analysing-microarray-data-in-bioconductor/

Exactly the same question has been asked elsewhere on this site:

  Annotating the results with associated gene symbols

The reason why Daniel Swan's code doesn't run exactly as is because the limma topTable now (in response to user requests) puts the probe IDs into the rownames of the output table rather than into a column called "ID". You can easily see this for yourself just by looking at the top table. So you just have to use rownames(probe.list) instead of probe.list$ID.

Alternatively you could follow the case study (Section 17.2) in the limma User's Guide, which is somewhat better because it puts the gene symbols in the data object right at the beginning of the analysis pipelline instead of the adding them later after the top table has been produced.