I am attempting a Golden Gate assembly with 4 pieces (including the backbone). I have sequenced each piece and verified the presence of BsaI sites on the end, with compatible overhangs. The overhangs chosen were screened using NEB's ligase fidelity tool. Using fresh enzyme and commercial grade competent cells, I got a bunch of transformants.
However, my results are confounding - I didn''t get the assembly I wanted across at least 12 different clones, but instead got some weird hybrids. Anyone have suggestions for improving / troubleshooting Golden Gate assembly?