The invention discloses a method for screening a controllable CcdB suicide switch based on
deep learning, and belongs to the technical field of
biology. According to the invention, a
deep learning tool MuToN
mutant and
DNA helicase binding energy change is used for prediction and sorting, mutants with obvious decline in affinity are screened, the obtained mutants are constructed into an expression
plasmid containing an inducible
promoter through
a site-specific
mutagenesis method, the expression
plasmid is transformed into an MG1655 bacterium lacking a detoxification
gene ccdA, and under the condition of not adding an
inducer, the expression
plasmid containing the inducible
promoter is transformed into the MG1655 bacterium lacking the detoxification
gene ccdA, and the expression plasmid containing the inducible
promoter is transformed into the MG1655 bacterium lacking the detoxification
gene ccdA. Screening a strain-free toxic
mutant through a
coating method; the obtained
mutant plasmid is transformed into MG1655
bacteria without a detoxification gene ccdA, a flat plate containing anhydrotetracycline is coated with the MG1655
bacteria, and a mutant with a high fatality rate is screened. A mutant which is moderate in
toxicity, can be expressed under the background of lacking CcdA and can reactivate
toxicity under an induction condition is constructed, so that a controllable suicide mechanism is realized, and the mutant can be used for on-demand self-killing of strains, construction of gRNA plasmids in a
CRISPR / Cas
system and sterile treatment of
fermentation waste liquid.