[BioC] simpleaffy parametres, Bioconductor Digest, Vol 59, Issue 15

[Patrícia Luiza Nunes da Costa]

Hi people

I changed the parameters of simpleaffy pairwise filter to:

min.exp=log2(10), min.exp.no=2, fc=log2(2), min.present.no=2, tt= 0.05,
present.by.group=TRUE

The situation didn’t change.
I had only 6 genes differentially expressed in the comparison of two
different groups. Each group has only 3 arrays.
Should I reduce the stringency? What parameter?

For others comparisons I had 50, 200 etc genes. But when I did the
clusterization
 everything was red!!!!
Anybody know what happened?

Thanks

Patrícia


> Hi Patricia
> your paramters are as follows
>
> min.exp - the minimum expression level you would typically expect to see
> in
> your experiment. For Affymetrix experiments, the default here has been set
> at log2(10), but that would obviously depend on your experiment and also
> how
> you have normalised the data
>
> min.exp.no - the minimum number of arrays you want your probeset to show
> the
> min.exp on. In your experiment, it is set at 2 so you  are filtering
> results
> to only show probesets with a minimum expression of log2(10) on at least
> two
> arrays
>
> min.present.no. - this is the minimum number of arrays a probeset must be
> called present on to be included in the analysis...in your case you have
> it
> set at 4.This means you are filtering your results to only include
> probesets
> that are present on at least 4 arrays.  Again the number here very much
> depends on your experiment, the number of groups you have and the number
> of
> arrays. In a simple experiment comparing 2 conditions using 3 arrays in
> each
> condition, I would set it at 2 or 3 (so the probeset has to be called
> present on half of the arrays in your experiment)
>
> present.by.group - this allows you to further specify how your present
> calls
> are. You have set it at FALSE, so your results are filtered to only show
> those probesets that are called present on at least 4 arrays in the whole
> experiment. If you set it to TRUE, then the probeset must be called
> present
> on at least 4 arrays in one of your conditions. Again what you set it at
> depends on your experiment. In my simple 6 array, 2 condition experiment I
> would set it to TRUE meaning my results will only include probesets that
> are
> present on at least 2 arrays in either condition. This allows you to
> account
> for probesets which may be absent in one condition, but present in another
>
> try typing openVignette() in your R session and selecting simpleaffy.
>
> hope this helps
>
> claire
>
> --------------------------
> Esta mensagem foi verificada
> pelo sistema de antivírus DIM e
> acredita-se estar livre de Virus.
>
>


Patrícia Luiza Nunes da Costa
Laboratório de Oncologia Experimental, Grupo de Adesão Celular
Faculdade de Medicina da Universidade de Paulo-FM USP
Av. Dr. Arnaldo, 455 sala 4112
Cerqueira Cesar
Cep 01246-903
Tel: (11) 3061-7486 e (11) 8202-7073



--------------------------
Esta mensagem foi verificada
pelo sistema de antivírus DIM e
acredita-se estar livre de Virus.