Rational selection of experimental readout and intervention sites for reducing uncertainties in computational model predictions

Table 2 Profile likelihood based identifiability analysis and confidence intervals of the in silico model example in log-space

Parameter	\({\hat {\theta }_{i}}\)	y = [D, A B] ^T			y= [D, A C] ^T			y =[D, B C] ^T			y =[D, A B C] ^T
		Identifiability	Lower CI	Upper CI	Identifiability	Lower CI	Upper CI	Identifiability	Lower CI	Upper CI	Identifiability	Lower CI	Upper CI
log10k ₁₁	0.043	identifiable	−0.72	0.44	non-identifiable	−∞	0.41	identifiable	−0.86	1.17	identifiable	−0.70	0.36
log10k ₁₂	0.301	non-identifiable	−∞	0.86	non-identifiable	−∞	∞	non-identifiable	−∞	0.82	non-identifiable	−∞	0.81
log10k ₂₁	0.398	identifiable	−0.16	0.60	identifiable	0.07	0.63	identifiable	−0.01	1.45	identifiable	0.12	0.59
log10k ₂₂	0.004	non-identifiable	−∞	3	non-identifiable	−∞	0.44	non-identifiable	−∞	0.39	non-identifiable	−∞	0.39
log10k ₂₃	−0.301	non-identifiable	−∞	∞	non-identifiable	−∞	0.31	non-identifiable	−∞	0.29	non-identifiable	−∞	0.23
log10d	0.004	non-identifiable	−∞	0.42	non-identifiable	−∞	0.40	non-identifiable	−∞	0.39	non-identifiable	−∞	0.37

This table illustrates the potential impact of additional readouts to the initial setup y = D.

ISSN: 1471-2105