probe_id function for affy data

Question

[akutasame]

Hello everyone,

I'm trying to fit a model to identify which genes are significant. But before model fitting I need to find a function for probe_id column, it should be able to determine how many probes there are and create a vector of that length, it needs to be flexible because some probeset are 1-16 or 1-11.

Thank you in advance

[Sean Davis]

It really sounds like you might be reinventing wheels here. Is there a reason not to follow what are now pretty standard approaches for dealing with microarray data normalization and summarization? Perhaps telling us what your data are and what you want to do, more generally, would be helpful.

[akutasame]

My affydata consists of three Cel files with no replicates(control,mutant and wildtype) I want to do a 2 way anova but before this I should find a factor that able to determine how many probes there are and create a vector of that length. Because some probeset are 1 to 16, 1 to 11 and 1-20.

"ProbesetID"      Control_HG-U133_Plus2.CEL                Mut-Sox2_HG-U133_Plus_2.CEL "           WT-Sox2_HG-U133_Plus_2.CEL"
"1007_s_at1"    517    468    586
"1007_s_at2"    2407    2182    2903
"1007_s_at3"    3161    3623    3826
"1007_s_at4"    6686    6286    7865
...
...
...

[Sean Davis]

I am still not sure exactly what you are looking to do, but this transformation (in R) might get you what you want. As an example:

sub('[0-9]+$','','1007_s_at12')
[1] "1007_s_at"

Now, you have actual Probeset IDs, with multiple rows for each probeset. That is, I suspect, the factor that you want, but I cannot really be sure.

[akutasame]

By the way, I did background correction with RMA method and normalization with Quantile method but not summarisation, i know it sounds like that but I m working according to what my supervisor wants from me

[akutasame]

Thanks for your help, But I m not sure if it is able to find how many probes there are? My main aim here is to fit an Anova with factor of probes and a factor of samples.

Apologies if this is a very basic question, I'm very new to this.

[Sean Davis]

The code I gave in my comment will generate a factor of probes (with multiple probes per probeset). If this isn't what you want, it would be helpful to clearly state what you are trying to do.

[akutasame]

Yes it is, just I couldn`t pass to two way Anova from here. I know the idea is like a column each row has the expression values, probe numbers and sample number(as factors); one column actual numbers, other column which probe is, other column which sample is.

[akutasame]

Any suggestion would be very useful, thanks