Bayesian DNA copy number analysis

BMC Bioinformatics

Table 2 Comparison among the several boundary estimators

	σ² = 0.1	σ² = 0.3	σ² = 0.5	σ² = 1.2
method	ρ² = 0.5	ρ² = 0.05	ρ² = 0.1	ρ² = 0.5
${\hat{T}}_{01}$ , ${\hat{ρ}}^{2}$	11.5967 ± 0.4562	18.0933 ± 0.6388	18.2 ± 0.6177	17.6833 ± 0.5781
${\hat{T}}_{j o i n t}$ , ${\hat{ρ}}^{2}$	8.8667 ± 0.3329	17.6333± 0.6224	17.4833± 0.5898	16.6167 ± 0.5367
${\hat{T}}_{B i n E r r}$ , ${\hat{ρ}}^{2}$	8.6833 ± 0.3437	17.3833 ± 0.6219	17.2467 ± 0.5983	15.8967 ± 0.5278
${\hat{T}}_{B i n E r r A k}$ , ${\hat{ρ}}^{2}$	8.7267 ± 0.3449	17.3933 ± 0.6226	17.2567 ± 0.5978	15.9133 ± 0.5303
${\hat{T}}_{01}$ , ${\hat{ρ}}_{1}^{2}$	11.4767 ± 0.4532	15.2867 ± 0.5325	15.6867 ± 0.536	16.1933 ± 0.5301
${\hat{T}}_{j o i n t}$ , ${\hat{ρ}}_{1}^{2}$	8.7933 ± 0.3294	13.73 ± 0.4731	13.81 ± 0.4691	14.1667 ± 0.451
${\hat{T}}_{B i n E r r}$ , ${\hat{ρ}}_{1}^{2}$	8.47 ± 0.3372	13.0567 ± 0.4635	13.18 ± 0.4725	13.2367 ± 0.4344
${\hat{T}}_{B i n E r r A k}$ , ${\hat{ρ}}_{1}^{2}$	8.4567 ± 0.3361	13.0467 ± 0.4676	13.08 ± 0.4695	13.2233 ± 0.4361

The table shows the estimated mean binary errors ± standard deviation of the estimators of t⁰ applied to datasets with replicates for different values of σ² and ρ². We always used ${\hat{K}}_{2}$ , $\hat{ν}$ , ${\hat{σ}}^{2}$ and both ρ² estimators as estimators of the other parameters involved. The estimators ${\hat{T}}_{B i n E r r}$ and ${\hat{T}}_{B i n E r r A k}$ always had the lowest errors. When σ² > ρ² the error was lower by using ${\hat{ρ}}_{1}^{2}$ , otherwise the two ρ² estimators gave similar errors.

ISSN: 1471-2105