Table 2 Attention map of the best model using the proposed method.

From: Subnetwork representation learning for discovering network biomarkers in predicting lymph node metastasis in early oral cancer

Subnetwork1

Attention2 (%)

Genes

EPITHELIAL_MESENCHYMAL

  

_TRANSITION_2

9.12

BMP1,DAB2,FBLN5,GADD45A,GEM,LOXL1,LUM,SNAI2,TPM1, VEGFA

E2F_TARGETS_5

9.06

AURKB,BRCA1,CCNE1,CDC20,CDKN2C,EXOSC8,GINS3,IPO7, MAD2L1,MCM7,POLA2,PRIM2,PTTG1,RAD1,RAD21,RANBP1, SYNCRIP,TK1,TUBB,XPO1

MYOGENESIS_6

7.97

ADAM12,CDKN1A,HSPB8,KIFC3,MB,MYOZ1,PSEN2,TNNC2, TNNT3,TPM3

TNFA_SIGNALING_VIA_NFKB_6

7.45

CCND1,CEBPB,CFLAR,ETS2,FOS,GADD45A,MAP3K8,MYC, NFE2L2,SMAD3,SPHK1,TNF,TRAF1,TRIB1

MITOTIC_SPINDLE_2

4.78

ARHGAP27,ARHGEF11,CENPE,CEP250,KIF4A,KIF5B,KIFAP3, LLGL1,LMNB1,RACGAP1,RASA1,TUBA4A

  1. Note that the top-5 subnetworks are listed.
  2. 1SUBNETWORK column indicates the name of subnetworks, where the prefix is the name of HGS geneset and the postfix (the integer number) is the index.
  3. 2ATTENTION column indicates the attention values generated by the best model of the proposed method averaged by all samples (including the TCGA and SNUH samples).
Back to article page