Figure 5

A Model of Mutagenic PCR. A model of a single cycle of mutagenic PCR. Each nucleotide of both sense and nonsense strands are treated as probability four-vectors. The 'state' of a nucleotide is the relative frequency we expect to observe it as either A, C, G, or T. The initial wild-type sequence is presumed to be well-defined. An example PCR cycle begins with A and T on the sense and nonsense strand, respectively, that are subsequently separated via denaturation. Error-prone polymerization of new nonsense and sense strands by Taq polymerase have, for this example, probabilities of 0.3 and 0.1 of nucleotide misincorporation. Random reassociation at the end of the cycle effectively averages the frequency of mutation at each site and implies that nucleotide frequencies are statistically independent despite being physically contiguous on the same strand. After the final PCR cycle completes a random sample of the DNA strands are selected for cloning, a fraction of which are subjected to selection based on the nucleotide sequence of the sense strand alone.