Skip to main content
. Author manuscript; available in PMC: 2016 Mar 1.
Published in final edited form as: J Am Stat Assoc. 2014 Oct 2;110(509):27–44. doi: 10.1080/01621459.2014.964404

Table 5.

The 95% interval of the sampling distribution of the posterior mean when experiment noise is correlated between genes. The results show that the estimate of ψ1,2 is essentially unbiased but the variation increases as the degree of between gene correlation increases. The fourth column shows the coverage of the 95% credible interval. While there is significant undercoverage, the error is small.

2.5% Mean 97.5% Coverage

ψ = 0.5, 10 blocks 0.46 0.50 0.52 0.81
ψ = 0.5, 100 blocks 0.47 0.50 0.52 0.91
ψ = 0.5, Brem et al 0.48 0.50 0.52 0.90

ψ = 0.8, 10 blocks 0.77 0.79 0.81 0.59
ψ = 0.8, 100 blocks 0.79 0.80 0.81 0.89
ψ = 0.8, Brem et al 0.78 0.80 0.82 0.74

ψ = 0.9, 10 blocks 0.86 0.89 0.91 0.28
ψ = 0.9, 100 blocks 0.89 0.90 0.91 0.76
ψ = 0.9, Brem et al 0.88 0.90 0.91 0.60