A numeric perturbation analysis revealed parameter value ranges that promote off-target effects in the n = 3 network. A perturbation analysis of the n = 3 network (G) was performed where a single parameter's value was randomly selected from a small range of values, while all other parameters were selected from the larger ranges defined in Table 1. The baseline in each plot reflects the percent of off-target effects in cycle 2 in 5000 sampled parameter sets when all parameter values were randomly selected from the ranges defined in Table 1 and depicted in (H). All other bars reflect the results of systematically perturbing each parameter (one at a time) using the given sub-ranges (A-F). Based on this perturbation analysis, a restricted parameter space was generated (I) from which ~74% of the sampled parameter sets produced off-target effects in cycle 2. In contrast, only ~1.6% of sampled parameter sets from the full parameter space (H) produced off-target effects in cycle 2.