Bayesian DNA copy number analysis

BMC Bioinformatics

Table 1 Comparison among the hyper-parameter estimators

	σ² = 0.1	σ² = 0.3	σ² = 0.5	σ² = 0.5	σ² = 0.5	σ² = 0.7	σ² = 1	σ² = 1.2
	ρ² = 0.5	ρ² = 0.05	ρ² = 0.02	ρ² = 0.05	ρ² = 0.1	ρ² = 0.5	ρ² = 0.05	ρ² = 0.5
${MSE}_{\hat{ν}}$	0.0904	0.0091	0.0059	0.0094	0.021	0.067	0.0169	0.0729
${MSE}_{{\hat{σ}}^{2}}$	0.0042	0.0014	0.0041	0.0036	0.0037	0.0114	0.0123	0.0272
${MSE}_{{\hat{ρ}}^{2}}$	0.0633	0.0871	0.2508	0.2404	0.2426	0.4271	0.9921	1.3254
${MSE}_{{\hat{ρ}}_{1}^{2}}$	0.068	0.0009	0.0008	0.0014	0.0047	0.0593	0.0024	0.0623

The table shows the estimated mean square error of the estimators $\hat{ν}$ , ${\hat{σ}}^{2}$ , ${\hat{ρ}}^{2}$ and ${\hat{ρ}}_{1}^{2}$ applied to datasets without replicates, for different values of σ² and ρ². The table shows that ${MSE}_{\hat{ν}}$ increased with ρ² and ${MSE}_{{\hat{σ}}^{2}}$ increased with σ². The estimator ${\hat{ρ}}_{1}^{2}$ was generally more accurate than ${\hat{ρ}}^{2}$ with respect to the MSE error measure.

ISSN: 1471-2105