[BioC] calculation of s0 (siggenes)

Holger Schwender holger.schw at gmx.de
Fri Oct 28 14:10:32 CEST 2005


Hi Assa,

the s0 value (s0=0) you get is the optimal value of s0 for your data. If you
prefer the original definition of Tusher et al. in which s0 is strictly
positive set include.zero=FALSE in sam(...). For more details on the
different options of sam(...) with the modified t-test see ?sam.dstat.

Best,
Holger


> --- Ursprüngliche Nachricht ---
> Von: Assa Yeroslaviz <Assa.Yeroslaviz at uni-duesseldorf.de>
> An: bioconductor at stat.math.ethz.ch
> Betreff: [BioC] calculation of s0 (siggenes)
> Datum: Fri, 28 Oct 2005 13:40:56 +0200
> 
> Hello,
> 
> IS there a way of calculating the s0-Value in the SAM analysis?
> I did a few tets with my data. Every time I get the result
> >data.sum
> ...
> s0 = 0
> 
>  Number of permutations: 200
> 
>  MEAN number of falsely called genes is computed.
> 
>  Delta: 9.3
>  cutlow: -23.092
>  cutup: 14.528
>  p0: 0.312
>  Significant Genes: 8
>  Falsely Called Genes: 0
>  FDR: 0
> 
> 
> Genes called significant (using Delta = 9.3):
> 
>     Row d.value stdev rawp q.value   R.fold      Name
> 1  4431    26.0  5.92    0       0 1.28e+42 254466_at
> 2 22511   -23.1 16.13    0       0 0.00e+00 259927_at
> 3  8199    18.4  5.61    0       0 1.83e+25 249456_at
> 4 14510    16.8  8.99    0       0 8.94e+40 258067_at
> 5  8351    16.4 10.68    0       0 3.08e+46 252077_at
> 6 16295    15.6  4.65    0       0 2.22e+20 263441_at
> 7  6168    14.9  7.62    0       0 1.35e+22 249015_at
> 8 22152    14.5 18.78    0       0      Inf 257088_at
> 
> I can define s0 as i like(in the sam line), but I'm not sure which value
> to
> give.
> 
> THX,
> 
> Assa
> 
> 
> --
> Assa Yeroslaviz
> Lötzenerstr. 15
> 51373 Leverkusen
> 
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> 

-- 
Telefonieren Sie schon oder sparen Sie noch?



More information about the Bioconductor mailing list