Table 3.
Comparison of results for Holm and Benjamini-Hochberg (BH) multiplicity corrections. SeaWAS results are given in columns labeled “NY”; replication results given in “C1,” “C2,” and “C3” columns.
| Total Number of Associations Identified | ||||
|---|---|---|---|---|
| NY | C1 | C2 | C3 | |
| Holm | - | 39 | 16 | 5 |
| BH | 55 | 81 | 46 | 9 |
| Number of Literature Supported Associations Identified [# (%Recall)] | ||||
| NY | C1 | C2 | C3 | |
| Holm | - | 9 (56) | 7 (44) | 1 (6) |
| BH | 7 (44) | 10 (63) | 11 (69) | 3 (19) |
| Number of Novel Associations Discovered | ||||
|---|---|---|---|---|
| NY | C1 | C2 | C3 | |
| Holm | - | 30 | 9 | 4 |
| BH | 16 | 71 | 35 | 6 |
| Number of Novel Associations Validated in Other Cohort [# (%Precision)] | ||||
| NY | C1 | C2 | C3 | |
| Holm | - | 6 (20) | 6 (67) | 3 (75) |
| BH | 4 (25) | 8 (11) | 14 (40) | 4 (67) |