Table 1.
Statistical information of the eight independent test datasets
| Independent test datasetsa | Positive | Negative |
|---|---|---|
| XUAMP | 1536 | 1536 |
| APD3 | 494 | 494 |
| DRAMP | 1408 | 1408 |
| LAMP | 1054 | 1054 |
| CAMP | 203 | 203 |
| dbAMP | 522 | 522 |
| YADAMP | 324 | 324 |
| DBAASP | 178 | 178 |
a
The relationship among the eight independent test datasets is shown in Figure 1. We adopt CD-HIT (Huang et al., 2010) to remove the redundancy between the independent datasets and the training sets. Following (Veltri et al., 2018), the similarity between positive training samples and positive independent test samples is <90%, and the similarity between negative training samples and negative independent test samples is <40%.