Table 2 Comparison of gene groups associated with signaling pathways

From: Inferring signaling pathway abnormalities from histopathological images via logic-constrained gene-pathway heterogeneous knowledge graph

Metric

Hamming loss

Mean ACC

Mean AUC

Backbones

Pathway

hsa04512

hsa04151

RHSA8948216

hsa04512

hsa04151

RHSA8948216

hsa04512

hsa04151

RHSA8948216

LCG-HGIN

0.1176

0.2429

0.1220

0.8824

0.7571

0.8780

0.8518

0.7214

0.8845

LCG-HGAT

0.0359

0.3235

0.1264

0.9641

0.6765

0.8736

0.9816

0.6851

0.9088

LCG-HSAGE

0.1618

0.2963

0.1983

0.8382

0.7037

0.8017

0.9071

0.6918

0.8952

LCG-HGCN

0.2010

0.2821

0.1852

0.7990

0.7179

0.8148

0.7781

0.6355

0.7646

SAGNN

–

–

–

0.7234

0.6832

0.6598

0.7499

0.7290

0.7183

CNN

–

–

–

0.5094

0.4923

0.5391

0.7705

0.7351

0.7697

AS_Graph

0.8333

0.2789

0.3593

0.1667

0.7211

0.6407

0.5000

0.6956

0.5000

AS_Weight

0.1797

0.2821

0.1438

0.8203

0.7179

0.8562

0.8133

0.7011

0.8740

DAMLN

0.1111

0.2196

0.1852

0.8889

0.8190

0.8148

0.6261

0.4789

0.4080

ADD-GCN

0.1552

0.4545

0.1667

0.4045

0.5455

0.4506

0.5250

0.5536

0.5509

Gene-DML

0.3333

0.2302

0.2407

0.6667

0.7698

0.7593

0.7623

0.7954

0.7942

  1. Bold values indicate the best performance achieved among all compared methods for each metric on the corresponding pathway.