BMC Bioinformatics

Table 1 Results of simulation study with linear mixed model

From: BICOSS: Bayesian iterative conditional stochastic search for GWAS

Setting	Measure	Method
Setting	Measure	SMA-exact	SMA-approx.	BICOSS	GWASinlps
	Recall	0.36	0.35	0.49	0.55
Setting 1	FDR	0.61	0.60	0.27	0.62
\(\beta ^{(1)} = 0.05\)	FPR \(\times 10 ^{5}\)	12.70	12.30	3.95	17.44
	F1	0.35	0.35	0.57	0.44
	Time (s)	197	2	22	85
	Recall	0.33	0.33	0.49	0.54
Setting 2	FDR	0.57	0.56	0.28	0.61
\(\beta ^{(1)} = 0.1\)	FPR \(\times 10 ^{5}\)	11.22	10.90	4.10	16.45
	F1	0.35	0.35	0.57	0.44
	Time (s)	203	2	46	122
	Recall	0.31	0.31	0.49	0.55
Setting 3	FDR	0.61	0.61	0.34	0.63
\(\beta ^{(1)} = 0.2\)	FPR \(\times 10 ^{5}\)	11.17	10.87	5.02	19.02
	F1	0.33	0.33	0.55	0.42
	Time (s)	201	2	47	117
	Recall	0.34	0.33	0.58	0.65
Setting 4	FDR	0.59	0.58	0.34	0.62
\(\beta ^{(1)} = 0.4\)	FPR \(\times 10 ^{5}\)	10.47	10.22	5.50	20.14
	F1	0.35	0.35	0.61	0.47
	Time (s)	203	2	50	130
	Recall	0.29	0.28	0.73	0.79
Setting 5	FDR	0.79	0.79	0.33	0.60
\(\beta ^{(1)} = 0.8\)	FPR \(\times 10 ^{5}\)	21.60	21.35	6.93	22.25
	F1	0.23	0.23	0.69	0.52
	Time (s)	186	1	44	148
	Recall	0.30	0.30	0.70	0.78
Setting 6	FDR	0.92	0.92	0.30	0.65
\(\beta ^{(1)} = 1.6\)	FPR \(\times 10 ^{5}\)	61.23	60.49	5.70	25.99
	F1	0.12	0.12	0.69	0.48
	Time (s)	176	1	45	147

Regression coefficients of causal SNPs \({\varvec{\beta }}= (\beta ^{(1)}, 0.4, 0.4, 0.4, \beta ^{(1)}, 0.4, 0.4, 0.4, \beta ^{(1)},0.4)^\top\)
Average Performance of each method over 100 datasets for each setting
Recall True Positive Rate, FDR False Discovery Rate, FPR False Positive Rate, F1 F1 score

Back to article page

ISSN: 1471-2105

Contact us

General enquiries: journalsubmissions@springernature.com