[BioC] Question on SAM output

[Holger Schwender]

Hi Alex,

this is explained in the R news article mentioned in the vignette of siggenes. I, however, will also add an explanation to this to the vignette.

False is *not* the number of False Positive, but what Tusher et al. call the number of falsely called genes. The number of False Positive is determined by p0 * False.

Best,
Holger


-------- Original-Nachricht --------
> Datum: Thu, 15 Nov 2007 15:31:43 -0500
> Von: "Alex Tsoi" <tsoi.teen at gmail.com>
> An: bioconductor at stat.math.ethz.ch
> Betreff: [BioC] Question on SAM output

> Dear all,
> 
> I used the sam analysis from the siggenes package to analyze
> differentially
> expressed genes, and here is the output:
> 
>   Delta    p0   False Called     FDR
> 1    0.1 0.235 3033.95   8740 0.08144
> 2    0.2 0.235 2539.17   8596 0.06930
> 3    0.3 0.235 2108.23   8452 0.05852
> 4    0.4 0.235 1788.62   8342 0.05030
> 5    0.5 0.235 1508.60   8226 0.04303
> 6    0.6 0.235 1261.23   8097 0.03655
> 7    0.7 0.235 1048.82   7964 0.03090
> 8    0.8 0.235  879.58   7828 0.02636
> 9    0.9 0.235  761.43   7719 0.02314
> 10   1.0 0.235  661.70   7580 0.02048
> 11   1.1 0.235  590.24   7476 0.01852
> 12   1.2 0.235  535.54   7407 0.01696
> 
> 
> I think I understand the mathematics and computation in SAM, but I am not
> quite sure how could I comprehend the results:
> The "False" column is supposed to be the false positive identified, and
> the
> "Called" column is the number of genes called to be significantly
> differentially expressed, so the FDR is calculated by using False to divid
> Called.
> 
> I understand the False is probably just from the average of the null
> distribution, but it still doesn't make sense:
> eg. if we look at the data when Delta is equal 0.2, there are arond
> 2539.17genes that are false positive, while the number of genes Called
> to be
> significant is 8596, then how could the FDR rate equals to 0.06930 ?
> 
> Am I understand this in a wrong way, could anyone please show me or direct
> me to the right direction ?
> 
> I greatly appreciate for any help or comment.
> 
> Thanks,
> Alex-
> 
> -- 
> Lam C. Tsoi (Alex)
> Medical University of South Carolina
> 
> 	[[alternative HTML version deleted]]
> 
> _______________________________________________
> 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

-- 
Ist Ihr Browser Vista-kompatibel? Jetzt die neuesten