TABLE 2.

The performance of models based on different negative data set constructions

Negative data set construction	5-fold cross-validationa		Independent testa
	AUC	AUPRC	AUC	AUPRC
Randomly selected genes	0.751 ± 0.003	0.554 ± 0.010	0.703 ± 0.013	0.483 ± 0.011
DAGs	0.662 ± 0.007	0.494 ± 0.011	0.552 ± 0.009	0.405 ± 0.011
HDFs from other viruses	0.625 ± 0.005	0.435 ± 0.008	0.539 ± 0.005	0.372 ± 0.005
Essential genes	0.703 ± 0.003	0.554 ± 0.010	0.762 ± 0.005	0.654 ± 0.012
Genes with similar T cell expression levels to HDFs	0.650 ± 0.002	0.461 ± 0.006	0.626 ± 0.009	0.441 ± 0.013
Genes with similar network degrees as HDFs	0.584 ± 0.003	0.399 ± 0.002	0.597 ± 0.002	0.401 ± 0.002

^aThe measurements are based on five different repeats of negative sample selections, which are reported as average ± SD.