Table 5 Cancer type experiment.

From: Multi-omics disease module detection with an explainable Greedy Decision Forest

Method

Selected genes

Accuracy

PPI distance

GO enrichment

[0,500,700.900]

p value

GDF (ours)

COG1, COG3, COG5

0.78

1,1,1,1

\(0.7^{-9}\)

GNN-SubNet

MGAT3, MGAT4B,MGAT5, MGAT5B

0.79

1,1,2,2

\(0.3^{-10}\)

RF

CUL2, LRSAM1, CCT3

0.81

2,2,2,2

\(0.1^{-2}\)

RF\(_{Boruta}\)

ATP2A1, CASP2, GMIP, HELQ, IQGAP3

0.81

2,4,4,5

\(0.1^{-2}\)

  1. Accuracy of the selected kidney cancer (KIRC) genes. GO term enriched genes are highlighted in bold. The PPI Distance reflects the longest shortest path from one node to the other, given a certain threshold for the confidence of the edges.
Back to article page