[BioC] Questions about multtest

[Rafael A. Irizarry]

On Tue, 24 Feb 2004, Oosting, J. (PATH) wrote:

<SNIP> 
> so I basically computed all p-values from the t.test also
> 
> mytt<-function(x,cl){ 
>   t.test(x[cl==0],x[cl==1])$p.val
> }
> 
> pvals<-apply(data[,6:12],1,mytt,cl= c(0,0,0,1,1,1,1))
> 
> Well, values turned out to be the same when doing this so I had to conclude
> there was no difference between my groups there

                                                                                
it should be pointed out that with just 3 reps the t-test is not a very
powerful test (low sensitivity). you might consider alternatives, such 
as the one performed by SAM (in the siggenes library) or one of the many 
implemented in limma. 

>
> 
> Keep in mind that the mt.maxT returns a sorted array with lowest p-values on
> top
> 
> Jan
> 
> > 
> > I run:
> > 
> > > t <- mt.teststat(data[,6:12], c(0,0,0,1,1,1,1), test="t")
> > 
> > which calculates the t statistic for my data.  The t 
> > statistic for my first gene comes up as:
> > 
> > > t[1]
> > [1] 40.60158
> > 
> > Presumably, this is equivalent to me running t.test:
> > 
> > > t.test(data[1,9:12], data[1,6:8], var.equal=FALSE, 
> > alternative="two.sided")
> > 
> >         Welch Two Sample t-test
> > 
> > data:  data[1, 9:12] and data[1, 6:8]
> > t = 40.6016, df = 2, p-value = 0.0006061
> > alternative hypothesis: true difference in means is not equal to 0
> > 95 percent confidence interval:
> >  1.713804 2.120092
> > sample estimates:
> >     mean of x     mean of y
> > -1.596190e-15 -1.916948e+00
> > 
> > So I want to know how I can get p-values for the t statistics 
> > I have just calculated using mt.teststat.
> >
> 
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor
>