Table 3.
Classifier performance in predicting GO terms in mouse, quantified by area under the ROC curve (AUC) and precision at 20% recall (P20R).
| Namespace | AUC | P20R | ||||
|---|---|---|---|---|---|---|
| MF | BP | CC | MF | BP | CC | |
| Cross-species | 0.90 | 0.67 | 0.81 | 0.52 | 0.16 | 0.42 |
| Species-specific | 0.86 | 0.83 | 0.86 | 0.42 | 0.29 | 0.46 |
| Multi-view | 0.91 | 0.81 | 0.88 | 0.57 | 0.30 | 0.58 |
| Chain | 0.89 | 0.82 | 0.87 | 0.51 | 0.28 | 0.52 |
The cross-species classifier uses only sequence data; the species-specific classifier uses a collection of genomic data--PPI, gene expression, and protein-GO term co-mention mined from the biomedical literature. The multi-view and chain classifiers are two approaches for integrating cross-species and species-specific data. The presented values are averages across all GO terms considered in a particular namespace. The results were obtained using five-fold cross-validation.