Table 1.

Proportions of simulated replicates for which the true underlying XCI model was identified by the max-LLR and comp-LLR approaches for SNP₁ and SNP₂.

Scenario	XCI model	MAF	LD	Max-LLRa		Comp-LLRb
				SNP1	SNP2	SNP1	SNP2
1	Skewed XCI to normal allele	0.3	0.95	0.828	0.840	0.828	0.840
2	Skewed XCI to deleterious allele	0.3	0.95	0.830	0.832	0.910	0.918
3	Random XCI	0.3	0.95	0.942	0.944	0.990	0.992
4	Escaping XCI	0.3	0.95	0.596	0.614	0.960	0.948
5	Skewed XCI to normal allele	0.4	0.95	0.970	0.974	0.970	0.974
6	Skewed XCI to deleterious allele	0.4	0.95	0.932	0.912	0.954	0.932
7	Random XCI	0.4	0.95	0.962	0.970	0.994	0.998
8	Escaping XCI	0.4	0.95	0.688	0.698	0.952	0.940
9	Skewed XCI to normal allele	0.3	0.85	0.616	0.660	0.618	0.660
10	Skewed XCI to deleterious allele	0.3	0.85	0.696	0.678	0.876	0.844
11	Random XCI	0.3	0.85	0.926	0.940	0.982	0.982
12	Escaping XCI	0.3	0.85	0.444	0.474	0.938	0.936
13	Skewed XCI to normal allele	0.4	0.85	0.860	0.834	0.864	0.834
14	Skewed XCI to deleterious allele	0.4	0.85	0.768	0.756	0.808	0.804
15	Random XCI	0.4	0.85	0.986	0.978	0.994	0.990
16	Escaping XCI	0.4	0.85	0.500	0.496	0.926	0.934

Abbreviations: LD: linkage disequilibrium (r²); LLR, likelihood ratio; MAF, minor allele frequency for the associated SNPs; SNP, single-nucleotide polymorphism; XCI, X-chromosome inactivation.

The proportions were calculated based on 500 replicates, each with 2000 cases and 2000 controls.

^aIf the maximum LLR corresponds to the true XCI model, we consider that the true model is identified using the max-LLR approach.

^bIf the true underlying XCI model is included in the resulting one or multiple equivalent possible XCI models, we consider that the true model is identified by the comp-LLR approach.