[BioC] background correction

[yanju]

Dear all,

As read from the instruction of backgroundCorrect in limma, the  method  
"half",  "minimum",  "edwards" and so on are all designed to produce 
positive corrected intensities.  My code is as follows:

    rgc<-backgroundCorrect(rg, method="half")
    plotDensities(rgc)

After I plot the graph of Density-Intensity after this backgroud 
correction, I found there is still some part of curve with the negative 
intensity. Why is it like that?? I try minimun also, the result is 
similar with negative intensity. I dont know how to explain it. 
Expecting your reply.

Regards,
Yanju