Salamander.

CV for I(tpeak) and Iarea by the FSR, TWS and GWS models. CVs computed for the WT and the following virtual conditions. LIN: i.e., the nonlinear fluxes Eqs 10–12 were linearized. NLD (non-local depletion): in the product [E*]σ[cGMP] in the last term on the right hand side of Eq 9, [cGMP] was kept constant at its dark value. LIN+NLD: both effects were enforced. Ca2+-clamp: Ca2+ dependence of cyclase activity was removed, i.e., in the expression of F in Eq 10, [Ca2+] was kept at its dark value [Ca2+]dark. In all cases and for FSR and and for TWS and GWS. The CVs of E* for all models converge to identical values with larger sample number.

针对I(tpeak)与Iarea，由FSR、TWS及GWS模型计算所得的变异系数（Coefficient of Variation, CV）。该变异系数针对野生型（Wild Type, WT）及下述虚拟实验条件进行计算：
LIN组：即对式10至12中的非线性通量实施线性化处理。
NLD（非局部耗竭，non-local depletion）组：在式9右侧最后一项的[E*]σ[cGMP]乘积项中，将[cGMP]固定为其黑暗条件下的数值。
LIN+NLD组：同时施加上述两种处理效应。
Ca²⁺钳制（Ca²⁺-clamp）组：移除环化酶活性对Ca²⁺的依赖性，即针对式10中F的表达式，将[Ca²⁺]固定为其黑暗条件下的数值[Ca²⁺]dark。
针对所有情况，FSR模型对应[原文缺失变量项]，TWS与GWS模型对应[原文缺失变量项]。
随着样本量增大，所有模型的E*的变异系数均收敛至相同数值。