Table 7 Results of the proposed model on different combinations of omics and networks at 500 genes using the TCGA pan-cancer dataset
From: A multimodal graph neural network framework for cancer molecular subtype classification
Data | Network | GAT | GCN | ||
|---|---|---|---|---|---|
Accu.e | F1 | Accu.e | F1 | ||
mRNAa | Intra-omicc | 77.0% ± 1.9% | 0.75 ± 0.03 | 76.1% ± 0.9% | 0.73 ± 0.01 |
miRNAb | Intra-omicd | 74.0% ± 0.4% | 0.70 ± 0.01 | 68.2% ± 4.1% | 0.63 ± 0.04 |
mRNA+CNVa | Intra-omicc | 79.1% ± 1.4% | 0.77 ± 0.03 | 77.1% ± 0.7% | 0.76 ± 0.01 |
mRNA+miRNA | Inter-omic | 76.1% ± 1.6% | 0.73 ± 0.03 | 75.4% ± 0.7% | 0.73 ± 0.01 |
Intra-omic | 77.3% ± 1.6% | 0.75 ± 0.03 | 76.8% ± 0.7% | 0.74 ± 0.01 | |
mRNA+CNV+miRNA | Inter-omic | 80.3% ± 1.6% | 0.80 ± 0.02 | 77.4% ± 0.6% | 0.74 ± 0.01 |
Intra-omic | 80.5% ± 1.2% | 0.80 ± 0.02 | 78.2% ± 0.6% | 0.75 ± 0.01 | |