Application of Bayesian network structure learning to identify causal variant SNPs from resequencing data

BMC Proceedings

Table 2 Primary results performed with log-likelihood scoring function on hill-climbing algorithm with 1,000 random restarts and 2,400 directional perturbations per score

Ancestry	Replicate	Descendants of Affected (DA)		Markov blanket of Affected (MBA)		Children of Affected (DA ∩ MBA)
		Number of causal SNPs/number of SNPs	P(X ≥ k)	Number of causal SNPs/number of SNPs	P(X ≥ k)	Number of causal SNPs/number of SNPs	P(X ≥ k)
Asian	1–10	12/25	0.0485	11/45	0.891	2/3	0.228
	11–20	2/7	0.695	9/36	0.850	0/2	1
	1–20	69/182	0.000467	15/55	0.798	2/7	0.695
	21–40	15/69	0.983	24/63	0.113	2/4	0.367
	1	77/237	0.152	63/221	0.973	16/54	0.658
	2	75/235	0.305	55/221	0.999	17/64	0.851
	3	74/217	0.0278	66/190	0.0355	22/43	0.00227
	4	76/227	0.0479	55/190	0.894	11/42	0.821
	5	76/240	0.371	66/218	0.758	25/68	0.155
	6	74/224	0.102	55/191	0.909	13/45	0.694
	7	67/203	0.160	60/202	0.826	16/55	0.693
	8	65/212	0.663	66/226	0.937	14/61	0.958
	9	73/230	0.368	67/226	0.887	24/86	0.816
	10	65/204	0.376	59/220	0.998	15/57	0.848
European	1–10	34/105	0.155	11/46	0.849	1/6	0.874
	11–20	6/11	0.0604	6/22	0.655	2/2	0.0822
	1–20	0/1	1	2/16	0.972	0/1	0.711
	21–40	35/107	0.124	9/34	0.703	3/4	0.0732
African	1–10	10/60	0.929	9/47	0.795	0/3	1
	11–20	22/99	0.613	10/44	0.566	0/4	1
	1–20	7/34	0.695	11/53	0.707	0/4	1
	21–40	2/2	0.0508	3/17	0.786	1/1	0.226

The analysis is for all typed, polymorphic SNPs in the target gene set. Probabilities less than 0.10 are shown in bold.

ISSN: 1753-6561