I am getting some rather odd "OddsRatio" numbers in my GOstat
tables. In particular, I seem to be getting "Inf" whenever
"Observed"="Size". This does not seem to be right, unless "Size" is
the number of hits in the union of the test and nontest set. But if
that is the case, then "Size" should always be greater than or equal
to "Observed" and that is not the case.
Naomi S. Altman 814-865-3791 (voice)
Associate Professor
Dept. of Statistics 814-863-7114 (fax)
Penn State University 814-865-1348
(Statistics)
University Park, PA 16802-2111
Hi Naomi,
Naomi Altman <naomi at="" stat.psu.edu=""> writes:
> I am getting some rather odd "OddsRatio" numbers in my GOstat
> tables. In particular, I seem to be getting "Inf" whenever
> "Observed"="Size". This does not seem to be right, unless "Size" is
> the number of hits in the union of the test and nontest set. But if
> that is the case, then "Size" should always be greater than or equal
> to "Observed" and that is not the case.
Hmm, I will certainly look into this. Can you provide some
sessionInfo() information?
The "Size" column is the number of genes annotated at the given GO
term (where genes are restricted to the defined gene universe). You
are right that Size should always be greater or equal to the observed
count. Jim found a bug relating to this which I fixed. If you don't
already have Category 2.0.3 and GOstats 2.0.4, could you try upgrading
(via biocLite) and let me know if you still see the problem?
+ seth
One thing I noticed is that the "universe" sent by the investigator
does not include all the genes in the test set. However, using the
union of the 2 sets as the universe does not fix the problem.
--Naomi
At 03:02 PM 12/11/2006, Seth Falcon wrote:
>Hi Naomi,
>
>Naomi Altman <naomi at="" stat.psu.edu=""> writes:
> > I am getting some rather odd "OddsRatio" numbers in my GOstat
> > tables. In particular, I seem to be getting "Inf" whenever
> > "Observed"="Size". This does not seem to be right, unless "Size"
is
> > the number of hits in the union of the test and nontest set. But
if
> > that is the case, then "Size" should always be greater than or
equal
> > to "Observed" and that is not the case.
>
>Hmm, I will certainly look into this. Can you provide some
>sessionInfo() information?
>
>The "Size" column is the number of genes annotated at the given GO
>term (where genes are restricted to the defined gene universe). You
>are right that Size should always be greater or equal to the observed
>count. Jim found a bug relating to this which I fixed. If you don't
>already have Category 2.0.3 and GOstats 2.0.4, could you try
upgrading
>(via biocLite) and let me know if you still see the problem?
>
>+ seth
>
>_______________________________________________
>Bioconductor mailing list
>Bioconductor at stat.math.ethz.ch
>https://stat.ethz.ch/mailman/listinfo/bioconductor
>Search the archives:
>http://news.gmane.org/gmane.science.biology.informatics.conductor
Naomi S. Altman 814-865-3791 (voice)
Associate Professor
Dept. of Statistics 814-863-7114 (fax)
Penn State University 814-865-1348
(Statistics)
University Park, PA 16802-2111
Naomi Altman wrote:
> I am getting some rather odd "OddsRatio" numbers in my GOstat
> tables. In particular, I seem to be getting "Inf" whenever
> "Observed"="Size". This does not seem to be right, unless "Size" is
> the number of hits in the union of the test and nontest set. But if
> that is the case, then "Size" should always be greater than or equal
> to "Observed" and that is not the case.
What is the output from sessionInfo()?
Best,
Jim
>
> Naomi S. Altman 814-865-3791 (voice)
> Associate Professor
> Dept. of Statistics 814-863-7114 (fax)
> Penn State University 814-865-1348
(Statistics)
> University Park, PA 16802-2111
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives:
http://news.gmane.org/gmane.science.biology.informatics.conductor
--
James W. MacDonald, M.S.
Biostatistician
Affymetrix and cDNA Microarray Core
University of Michigan Cancer Center
1500 E. Medical Center Drive
7410 CCGC
Ann Arbor MI 48109
734-647-5623
**********************************************************
Electronic Mail is not secure, may not be read every day, and should
not be used for urgent or sensitive issues.
Just a quick update:
Naomi sent me data offlist and I can reproduce the problem with the
count and size. I'm investigating...
I think it may be related to the ath1121501 annotation package and
differences in how it is structured from other annotation packages.
> One thing I noticed is that the "universe" sent by the investigator
> does not include all the genes in the test set.
Clearly that doesn't make a lot of sense. I'm pretty sure the
hyperGTest code would remove the genes from the test set that are not
in the specified universe.
+ seth
