Protein Property Prediction on 21 protein landscapes 1536 train points (cross-val)
0.867Spearman CorrelationLOCK-GP
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| LOCK-GP2026.06 | 0.867 | 0.914 | 0.21 | |
| ConFit2026.06 | 0.863 | 0.867 | 0.344 | |
| MLP-ESM2-8MBackbone=ESM2-8M2026.06 | 0.856 | 0.898 | 0.256 | |
| MLP-ESM2-LastLayerBackbone=ESM2, Pooling=LastLayer2026.06 | 0.851 | 0.9 | 0.243 | |
| MLP-SaProt-LastLayerBackbone=SaProt, Pooling=LastLayer2026.06 | 0.851 | 0.903 | 0.238 | |
| Kermut-GP2026.06 | 0.85 | 0.888 | 0.285 | |
| Tanimoto-GP2026.06 | 0.846 | 0.888 | 0.272 | |
| MLP-ESM2-8M-RandInitBackbone=ESM2-8M, Initialization=Random2026.06 | 0.839 | 0.88 | 0.295 | |
| Ridge-ESM2-8MBackbone=ESM2-8M, Classifier=Ridge2026.06 | 0.837 | 0.877 | 0.301 | |
| Ridge-ESM2-650MBackbone=ESM2-650M, Classifier=Ridge2026.06 | 0.836 | 0.877 | 0.304 | |
| MLP-ESM2-650MBackbone=ESM2-650M2026.06 | 0.828 | 0.873 | 0.288 | |
| KermutStruc-GP2026.06 | 0.809 | 0.805 | 0.417 | |
| Ridge-OH-ESM2-650M-AugFeatures=One-Hot + ESM2-650M-Aug, Classifier=Ridge2026.06 | 0.804 | 0.805 | 0.419 | |
| SigGLM-OHBackbone=One-Hot2026.06 | 0.804 | 0.849 | 0.342 | |
| Ridge-OHBackbone=One-Hot, Classifier=Ridge2026.06 | 0.802 | 0.8 | 0.422 | |
| MLP-ESM2-8M-MeanPoolBackbone=ESM2-8M, Pooling=MeanPool2026.06 | 0.797 | 0.843 | 0.362 | |
| KermutSeq-GP2026.06 | 0.794 | 0.838 | 0.338 | |
| Ridge-ESM2-8M-MeanPoolBackbone=ESM2-8M, Pooling=MeanPool, Classifier=Ridge2026.06 | 0.786 | 0.829 | 0.376 | |
| MLP-OH-ESM2-650M-AugFeatures=One-Hot + ESM2-650M-Aug2026.06 | 0.771 | 0.818 | 0.442 | |
| MLP-OHBackbone=One-Hot2026.06 | 0.764 | 0.808 | 0.459 | |
| BLOSUM50-ZeroShotMode=ZeroShot2026.06 | 0.396 | 0.386 | — | |
| ESM2-650M-ZeroShotMode=ZeroShot2026.06 | -0.044 | -0.04 | — |