[BioC] Controls and duplicate correlation

[lepalmer at notes.cc.sunysb.edu]

I was searching the archives for the warning message I got with 
duplicateCorrelation.  I got that resolved but found this comment from 
Gordon from last November.

"Secondly, assigning zero weight is not the way to remove missing or 
control spots. You should 
leave the weights as they are are and instead subject the data object. For 
example, 
  isBlank <- (MA$genes$Status %in% c("blank","miss")) 
  corfit <- duplicateCorrelation(MA[!isBlank,], design, ndups=1, 
block=pair) 
See the Weaver example in the Limma User's Guide for an example of the 
treatment of control spots. "

I looked at the Weaver example, and wasnt sure what this last line 
referred to.

I tried doing this to my data:

isBlank <- (MA$genes$Status %in% c("Blank","Reserved"))
corfit <- duplicateCorrelation(MA[!isBlank,], design, ndups=2,spacing=240)

but got this error
Error in unwrapdups(M, ndups = ndups, spacing = spacing) : 
        dim<- : dims [product 57600] do not match the length of object 
[57948]

Summary of isBlank
   Mode   FALSE    TRUE 
logical    9658    1862 

The data consists of 3 sets of dye swaps, with 5760 spots printed twice on 
chip (11520 spots total).

There are 'Blanks' (nothing printed) and 'Reserved' which is DNA from 
another species which should be expressed.

These blanks and reserved are flagged in .gpr file and should have a 
weight of 0.

My questions are:
What am I doing wrong?
Is it still necessary to do above to remove blanks and reserved?
Is this removal of control spots required at any other time?

I like having the reserved spots being analyzed, as they should not show 
up in statistically significant genes.  If they do, then I know that I did 
not do the calculations correctly.  Is it wrong to leave these in with the 
analysis (they are still weighted at 0)

Thanks,
Lance Palmer





	[[alternative HTML version deleted]]