Association analyses of the MAS-QTL data set using grammar, principal components and Bayesian network methodologies

Table 1 Comparison of linkage disequilibrium measures with general Bayesian Network arc strengths.

Marker1	Marker2	Chi	P(Chi)	D	CorrCoeff	Dprime	Delta	PropDiff	YulesQ	ARC	exp(ARC)
A8111	A9100	692.12	1.50x10^-152	0.09	0.55	0.87	0.88	0.71	0.95	293.53	3.01x10¹²⁷
A8363	A9100	516.77	2.10x10^-114	0.08	0.47	0.99	0.99	0.66	0.99	71.68	1.35x10³¹
A8111	A8363	548.72	2.40 x10^-114	0.11	0.49	0.64	0.51	0.45	0.82	519.43	3.85x10²²⁵
A8111	A8351	2318	1.70 x10^-236	0.16	0.68	0.98	0.63	0.63	0.99	694.50	4.14x10³⁰¹
A8035	A8329	1668.98	0	0.21	0.85	0.97	0.84	0.83	1.00	232.66	1.10x10¹⁰¹
A8329	A8351	20.12	7.27 x10^-6	-0.02	-0.09	-0.11	-0.19	-0.09	-0.19	240.85	3.98x10¹⁰⁴

Arc (and it is exponent) shows that taking the arc away from the current network would make the resulting model less probable; hence bigger arc number shows stronger association.
D Linkage Disequilibrium Coefficient
CorrCoeff: Correlation coefficient
Dprime: Lewontin’s D’
Delta: Population attributable risk, δ
PropDiff: Proportional difference
YulesQ: Yule’s Q

ISSN: 1753-6561