Skip to main content

Table 1 Permeability prediction results under ID setting

From: MultiCycPermea: accurate and interpretable prediction of cyclic peptide permeability using a multimodal image-sequence model

Model

MSE (↓)

MAE (↓)

R2 (↑)

PCC (↑)

SCC (↑)

SVM

0.48

0.48

0.18

0.46

0.46

RandomForest

0.30

0.38

0.48

0.70

0.66

TextCNN

0.59

0.57

0.00

0.07

0.12

BiLSTM

0.59

0.57

0.00

 − 0.05

 − 0.05

GAT

0.47

0.50

0.19

0.45

0.46

GCN

0.49

0.52

0.17

0.41

0.43

ChemBERTa

0.26

0.38

0.61

0.76

0.69

Unimol

0.24

0.36

0.59

0.77

0.72

MolCLR

0.19

0.33

0.67

0.82

0.76

CyclePermea

0.22

0.34

0.63

0.81

0.74

Multi_CycGT

0.29

0.40

0.45

0.71

0.68

MultiCycPermea

0.16

0.28

0.73

0.85

0.81

Back to article page

BMC Biology

ISSN: 1741-7007

Contact us