Assessing the effect of genetic markers on drug immunogenicity from a mechanistic model-based approach

Table 2 Power of the TWIST compared to the TLRT when α and γ are additive

	Underdispersion		Equidispersion		Overdispersion
	(γ₁**=−0.35andγ₂=0)**		(γ₁=0andγ₂*=0)*		(γ₁**=0.10andγ₂=0)**
e^β	TWIST	TLRT	TWIST	TLRT	TWIST	TLRT
(a) α has an additive effect and e^α>1
2	0.95	0.56	0.42	0.08	0.49	0.42
1	0.93	0.36	0.29	0.02	0.37	0.28
0.50	0.95	0.47	0.32	0.03	0.42	0.32
(b) α has no effect, e^α=1
2	0.50	0.44	0.09	0.10	0.59	0.60
1	0.36	0.25	0.04	0.02	0.50	0.48
0.50	0.45	0.34	0.04	0.04	0.53	0.53
(c) α has an additive effect and e^α<1
2	0.22	0.32	0.41	0.12	0.95	0.77
1	0.11	0.14	0.30	0.04	0.91	0.65
0.50	0.14	0.18	0.35	0.05	0.96	0.74

(b) The underlined values show the estimated level of the type I error.
(c) The tail defects for the [AA], [Aa] and [aa] groups are respectively : 0.50, 0.55 and 0.59 for γ₁= -0.35, and 0.50, 0.44 and 0.37 for γ₁= 0.10. The proportion of pre-immune subjects is 10% and the MAF is 20%.

ISSN: 1471-2288