Table 3 Parameter estimates (true model: convolution of normal and log-normal distribution)
From: Modeling gene expression measurement error: a quasi-likelihood approach
| Â | EQL | ANL | true value | ||||
|---|---|---|---|---|---|---|---|
Replicates | 4 | 10 | 20 | 4 | 10 | 20 | Â |
err( | 0.0008 | 0.0007 | 0.0006 | 0.0010 | 0.0014 | 0.0009 | 0 |
| 20605 | 20012 | 19332 | 22979 | 24178 | 25161 | 25000 |
| 0.1906 | 0.2035 | 0.2032 | 0.2340 | 0.2395 | 0.2427 | 0.25 |
| 4150.1 | 4559.8 | 4669.6 | 4995.4 | 5053.6 | 5082.3 | 5000 |
-log L | 280822 | 708618 | 1420480 | 280205 | 708308 | 1419534 | Â |
i
)
- err(
i
) = avg((
i
- μ
i
)/μ
i
) EQL: all parameters estimated via EQL ANL: all parameters estimated using asinh-normal assumption