This is a statistics toy about genome editing, not a protocol. Each cut attempt by a guided nuclease ends in one of three outcomes: it lands off-target (probability 1 - P), it cuts the target but the cell repairs it unchanged (P(1 - E)), or the intended edit is made (P · E), where P is precision and E efficiency. So the success rate is P · E, and the number of edits after n attempts is Binomial(n, P E).
If a site is attacked m times, the chance it is edited at least once is 1 - (1 - P E)m; the readout compares the sampled number of edited sites with that expectation. Time runs at 1 simulated minute per second with the original page's toy rates: 3 cut attempts, 2 ligations and 5 fragments per minute.
Simplifications: the technique selector only changes the colour theme and the one-line description, the same probability model is used for all four; real editing efficiencies depend on the cell type, target and delivery; the protein and fragments are stylised shapes, not molecular structures; real off-target cuts are rare and are checked with sequencing. Real gene editing is done only under licensed laboratory oversight, never at home.
Running numerical checks...