Table 3. Parameter determinability in mRNA- and protein-based gene circuits.

Description of Scenario	Data	Scoring Function	Weights for WLS	Trunk Region (Nuclei)	Independent Confidence Intervals	Determinability
Gene circuits using protein data from [49]	protein	WLS	derived from protein data	58	20/5/4	good
Gene circuits using protein data and mRNA-style weights	protein	WLS	mRNA-style	58	16/6/7	reasonable
Gene circuits using protein data approximated by mRNA-style boundary extraction	protein	WLS	mRNA-style	58	11/3/15	reasonable
Gene circuits from mRNA and WLS, with 58 nuclei	mRNA	WLS	mRNA-style	58	0/3/26	poor
Gene circuits from mRNA and WLS	mRNA	WLS	mRNA-style	53	0/2/27	poor
Gene circuits from mRNA and OLS	mRNA	OLS	not used	53	0/1/28	poor

Each row represents the results of a series of optimisation runs to data described in columns 2–5: mRNA- or protein-based fits, OLS or WLS cost function, variance-based or approximated (mRNA-style) weights for WLS, and region covered by models (53 or 58 nuclei). Column 6 (‘Independent Confidence Intervals’) shows triplets, which represent the number of regulatory parameters in fitted models that are determinable/weakly determinable/non-determinable. Determinable parameters are those whose confidence intervals fall exclusively into one regulatory category (activating, no interaction, or repressing). Weakly determinable parameters are those where one regulatory category is excluded from the confidence interval (‘not repressing’, or ‘not activating’). Confidence intervals for regulatory weights in all scenarios are shown in Text S4. Overall determinability of parameters is summarised in column 7.