[BioC] DESeq no recognizing row.names

[Steve Lianoglou]

Hi Alicia,

Unfortunately I replied to the previous email off list, but let's get back on list now:

> Sorry for my stupidity, i'm completely new in R and DESeq.
> 
> I identified the problem but because i opened my file in excel and i saw that i have some decimal numbers. But i didn't identified at all my problem by what i got with which( spercysts_vs_embryos != round (spercysts_vs_embryos),
> >      arr.ind=TRUE )
> 
> Can you please explain to me what i got?:
> 
> comp203811_c0_seq3    44259   1
> > comp203811_c0_seq4    44260   1
> > comp203818_c0_seq2    44266   1
> > comp203818_c0_seq4    44268   1
> > comp203827_c0_seq2    44281   1
> > comp203827_c0_seq3    44282   1
> > comp203827_c0_seq7    44286   1
> > comp203828_c0_seq1    44287   1
> >  [ reached getOption("max.print") -- omitted 166743 rows ]

You've just identified the row,col indices of the values in your count matrix that do not look like integers.

As Simon and I tried to explain in the initial responses, this command:

R> which( spercysts_vs_embryos != round (spercysts_vs_embryos), arr.ind=TRUE )

Finds the elements in `spercysts_vs_embryos` that are not rounding to each other. As you included in the previous email, one of these elements is (44259,1), so what is it?

R> spercysts_vs_embryos[44259, 1]

The point is that DESeq needs *integers*, you have stuff in your data.frame that are not integers. Likely they are some decimal number, I guess. You need to input raw count data to DESeq, not some mean count, RPKM, or whatever.

HTH,
-steve

--
Steve Lianoglou
Computational Biologist
Bioinformatics and Computational Biology
Genentech