Hi Heyi,
We considered this scenario with argument 'weights' in the gage
function. To use this argument, we need to manually pair arguments
'ref' and 'samp', and set compare='paired'. Check the help information
for the function by typing '?gage' within R. take a simple example, we
have 7 samples/columns C1a, C1b, C2a, C2b, T1a, T1b, T2a (T2b is
missing somehow). C means control, T means treated, 1 and 2 are
biological replicates, a and b are technical replicates. When C and T
samples are matched, i.e. C1-T1, C2-T2, we set: ref=c(1,1,2,2,3,4),
samp=c(5,6,5,6,7,7), weights=c(0.5,0.5,0.5,0.5,1,1), compare='paired'.
When C and T samples are unmatched, ref=rep(1:4, each=3),
samp=rep(5:7, 4), weights=rep(c(0.5, 0.5, 1), 4), compare='paired'.
There is no need to set weights option if T2b exists, as every C and T
sample has the same weight. Hope this helps.
Weijun
--- On Fri, 12/3/10, heyi xiao <xiaoheyiyh at="" yahoo.com=""> wrote:
> From: heyi xiao <xiaoheyiyh at="" yahoo.com="">
> Subject: gage package: biological replicates and technical
replicates
> To: bioconductor at stat.math.ethz.ch
> Date: Friday, December 3, 2010, 4:44 PM
> I am doing gene set analysis using
> gage package. I have two sample groups, controls and treated
> samples. I have one question about using the default 1-on-1
> (pairwise) comparison for unpaired samples. My samples have
> both biological replicates and technical replicates. Shall I
> treat them the same way? Woud that be a little improper?
>
> Heyi
>
>
> ? ? ?
>
