Table 3.

MAE and CC comparison of different methods in the prediction of GTEx data

Methods	MAE	CC
LSR	0.4704 ± 0.1235	0.7184 ± 0.2072
LSR-L1	0.5669 ± 0.1274	0.6813 ± 0.2188
LSR-L2	0.4682 ± 0.1233	0.7181 ± 0.2076
KNN-GE	0.6520 ± 0.0982	0.3941 ± 0.4124
D-GEX	0.4393 ± 0.1239${}^{\text{⋆}}$ / 0.4380 ± 0.1237	− / 0.7337 ± 0.2072
GGAN	0.4215 ± 0.1264	0.7475 ± 0.2070

Note: The results of D-GEX is based on a network with two hidden layers and 9000 hidden nodes in each layer (best structure reported). The results of GGAN is based on the DenseNet architecture. The results of the comparing models are obtained by us running the released codes, except the one marked by ($\text{⋆}$) on top that is reported from the original paper. Better results correspond to lower MAE value or higher CC value