The quality of fit of the model usually improves in successive cycles through the Computing Life paradigm. Several measures of fit are reported.

The number of data points (n) used in fitting, χ2 goodness of fit measure [7], which is cumulative across cycles, estimates of the error variance σ2, Gene Expression Level 50 (GEL50)* as a proxy for power [77], and the fold-change in expression level corresponding to the GEL50 are reported to allow comparison with future and existing models and microarray experiments. The genetic network fitted is shown in Fig. 4 (except cycle 4).
*The gene expression level 50 (GEL50) was the median value of the test statistic among genes with a significant F or t statistic for a circadian, light-, or WCC-response and with LREs upstream. This measure is an indicator of power [77] to detect one of these three responses, allowing comparisons with other microarray experiments. The fold-change in expression level (e.g., D/L) corresponding to the reported GEL50 is reported in the last column when applicable.
$The estimated error variance was computed from the number of observations (n), the preliminary estimate of σ02 = 0.02 [7], and the χ2min over the ensemble using the formula: .
%The square-root of this F1,9-statistic is reported using the fact that F1,9 = t92 to allow a comparison with other t-statistics in the same column.