BAC recombineering problem - BAC clones positive for double selection but negative for PCR, why? (Dec/02/2010 )

I have been trying to use the BAC recombineering strategy to insert a special cassette (for tracking in human embryonic stem cells) into the target region of the gene of interest by homology (70bp by PCR or 500bp by cloning). However, after trying for several months by various approaches, I always seem to get colonies after the targeting that are positive for cassette selection (kan resistance)and BAC selection (chloramphenicol resistance) suggesting successful targeting but when I do a PCR with primers flanking the cassette, I always get predominantly a smaller band indicative of no cassette insertion and maybe a faint band with a size suggesting incorporation of cassette. I somehow cannot get a pure colony with just the recombined vector. I tried retransforming, linearizing-ligating-retransforming but always get the same pattern after PCR. There is no significant homology of cassette to the bacterial genome (the cassette has a Neo part which has not been modified) to rule out recombination with bacterial genome. Has anyone come across this kind of problem before and do you have any suggestions how to solve this problem please? I am desperate now!
I am using the SW106 recombinogenic strain and follow the protocol of recombination (from NCI-frederick)to the word!

Thanks a lot

hey there! I know this is kind of a late reply. I am hoping by now you would have found your solution! Because I am facing the same problem. What I am doing is recombineering a Kanamycin resistance cassette with 60bp homology into pBeloBAC11 plasmid. And I almost always get colonies growing on LB + chloramphenicol + kanamycin. But whenever I do PCR screening, the non-recombinant band (shorter band signifying no insertion) is always stronger than the faint recombinant band. I tried altering the PCR conditions such as using significantly less primers and increasing the annealing temperature, which successfully increase the intensity of the long recombinant band. However, the non-recombinant band is equally strong as well.