To generate stable mammalian cell lines, I am doing a kill curve to find the optimal concentration of antibiotic for selection. The problem I have is that, even with a very high level of antibiotic (~10X common literature values), a very low % of cells look kind of "okay" by light microscopy. Some protocols seem to indicate a low % survival is okay (eg, only 95% killing is required), other protocols say that you must kill 100% of cells.
My question is: What is the criteria for the "minimum" selection concentration for stable cell lines?
1. 100% of cells dead?
2. 95% of cells dead?
3. No growth of culture after a week?
4. Some other requirement?