[BioC] is limma robust to NA values?

[magates at u.washington.edu]

Hello, 
 
For ease of compatiblity with other manipulations I do, I would like to set  
to "NA" those R,G,Rb,Gb, M,A values for which my weight is "0".  Is limma 
robust to having NA values present, including rows/genes where all values are 
NA? 
I ask because a few of the other array packages I use are rather cranky about 
NA values.  
  
Thanks,   
Michael Gates