TABLE 1.
Results from a simulation study on regulator–target pair identification in a single-eQTL region with three causal polymorphisms and with multiple candidate regulator and target genes (true network structure is in Figure 1)
| Methods | SD = 0.5 | SD = 0.25 | SD = 0.125 | SD = 0.5/0.125 | SD = 0.5/0.25 | SD = 0.25/0.125 |
|---|---|---|---|---|---|---|
| Cis-link, power (%) | 100, 100 | 100, 100 | 100, 100 | 55.3, 59.85 | 89.4, 98.65 | 97.8, 98.5 |
| Cis-link, FDR (%) | 0.6, 0.9 | 0.7, 0.67 | 0.67, 0.57 | 0.48, 0.53 | 0.6, 0.72 | 0.62, 0.57 |
| Cis-reg cis, power (%) | 99, 99 | 99, 99 | 99, 99 | 54.8, 59.4 | 88.6, 97.8 | 97.2, 97.8 |
| Cis-reg cis, FDR (%) | 0.35, 0.13 | 0, 0 | 0, 0 | 38.6, 0.25 | 3.27, 0.15 | 1.8, 0.1 |
| Cis-reg, power (%) | 99, 98.8 | 99, 98.9 | 98, 98.5 | 54.9, 59.2 | 88.2, 97.4 | 96.9, 97.5 |
| Cis-reg, FDR (%) | 0.93, 0.4 | 0, 0 | 0, 0 | 45, 25.82 | 4.82, 1.33 | 2.85, 1.3 |
| Trans-reg, power (%) | 99, 99.2 | 100, 100 | 100, 100 | 41.8, 45.1 | 92.9, 96.1 | 96.3, 96.5 |
| Trans-reg, FDR (%) | 0.85, 1.1 | 1.1, 1.15 | 1.57, 1.52 | 26.95, 62.52 | 10.92, 2.52 | 2.3, 2.77 |
Power, percentage of replicate data sets in which the regulation type was found; FDR, percentage of replicate data sets in which a regulation of a certain type was found that did not exist in the underlying network; Cis-link, cis-regulation of target in eQTL region; Cis-reg, cis–trans-regulation of target not in eQTL region; Cis-reg cis, cis–trans-regulation of target in eQTL region; Trans-reg, trans-regulation. For the last three columns, even-numbered gene nodes (Figure 1) received the left amount of error variance and odd-numbered nodes the right amount. The two numbers in each cell correspond to 0% recombination and 9% recombination among the three causal polymorphisms in the single-eQTL region, respectively. A P-value cutoff of 0.01 was used.