Detection of Fetomaternal Genotype Associations in Early-Onset Disorders: Evaluation of Different Methods and Their Application to Childhood Leukemia

Table 1

Forward stepwise likelihood-ratio testing procedure used to dissect child and maternal genotype associations.

Weinberg and Umbach’s log-linear approach using case-triads, unrelated cases and unrelated controls (HD-NPC) or parents of controls (HD)	Genotypic case-control test combined with the conditional logistic regression approach of Cordell et al. using case-triads (CC+CLR)
Step 1.	Step 1.
Child genotypic effect	Child genotypic association test (CC)
Null versus CG  LR Chi-square (2df)	Chi-square or Fisher’s exact test (2df)
And	And
Maternal genotypic effect	Maternal genotypic effect (CLR)
Null versus MG  LR Chi-square (2df)	Null versus MG  LR Chi-square (2df)
Step 2. (depending on Step 1)	Step 2. (depending on Step 1)
Maternal effect given child effect	Maternal effect given child effect (CLR)
CG versus CG+MG  LR Chi-square (2df)	CG versus CG+MG  LR Chi-square (2df)
Or	Or
Child effect given maternal effect	Child effect given maternal effect (CLR)
MG versus CG+MG  LR Chi-square (2df)	MG versus CG+MG  LR Chi-square (2df)

Likelihood-ratio tests were performed in a forward stepwise fashion. The most significant single-step test (Child versus Null or Mother versus Null) was tested against a joint effects model in a 2 degree-of-freedom likelihood-ratio test (Child + Mother versus Child or Child + Mother versus Mother). LR Chi2 indicates likelihood-ratio chi-square test; df, degrees of freedom; GC Child genotype relative risk; GM, Mother genotype relative risk.