Table 3 Permeability prediction results under Cliff setting

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

Model	MSE (↓)	MAE (↓)	R² (↑)	PCC (↑)	SCC (↑)
SVM	0.47	0.58	0.16	0.64	0.63
RandomForest	0.30	0.40	0.57	0.76	0.72
TextCNN	0.75	0.66	0.00	0.10	0.11
BiLSTM	0.75	0.66	− 0.02	0.00	− 0.04
GAT	0.56	0.56	0.23	0.48	0.50
GCN	0.62	0.58	0.15	0.41	0.45
ChemBERTa	0.29	0.39	0.60	0.79	0.73
Unimol	0.32	0.41	0.58	0.69	0.64
MolCLR	0.23	0.35	0.71	0.84	0.80
CyclePermea	0.27	0.38	0.63	0.80	0.72
Multi_CycGT	0.31	0.41	0.56	0.73	0.65
MultiCycPermea	0.18	0.29	0.76	0.88	0.85

Back to article page

ISSN: 1741-7007

Contact us

General enquiries: journalsubmissions@springernature.com