Table 1.

Mean term-centric ROCAUC achieved by the methods under comparison using 3-fold cross-validation (CV, second and third column) and when testing on the CAFA3 dataset (CAFA3, fourth and fifth column)

Method	ROCAUC (CV)	Weighted ROC AUC (CV)	ROC AUC (CAFA3)	Weighted ROCAUC (CAFA3)
PCC	0.69 ± 0.003	0.69 ± 0.003	0.68 [0.63, 0.72]	0.68 [0.63, 0.73]
PCC(k)	0.69 ± 0.003	0.69 ± 0.003	0.68 [0.63, 0.71]	0.68 [0.63, 0.72]
PCC + MR	0.72 ± 0.002	0.72 ± 0.002	0.69 [0.65, 0.73]	0.69 [0.65, 0.73]
GAAWGEFA	0.71 ± 0.002	0.71 ± 0.002	0.69 [0.65, 0.73]	0.70 [0.65, 0.74]
MLC ($S_{}w$)	0.72 ± 0.003	0.73 ± 0.003	0.69 [0.65, 0.73]	0.69 [0.65, 0.73]
MLCG	0.72 ± 0.003	0.72 ± 0.003	0.71 [0.67, 0.75]	0.72 [0.67, 0.76]
MLC-MR Hybrid	0.73 ± 0.005	0.73 ± 0.005	0.69 [0.65, 0.73]	0.69 [0.66, 0.73]

Notes: For the cross-validation, we report the average performance over the three folds as well as the corresponding standard error. For the CAFA3 results we report the performance on the test set as well as the 95% confidence intervals from doing 1000 bootstrapped tests. The top performance of every column is shown in bold.