Leaderboard

Benchmarking state-of-the-art models on StereoPep. Results are reported as mean ± std over multiple runs.

Want to add your model? Open a pull request on GitHub with your results.

Retention Time Prediction — Regression

Mean ± std over seeds. Bold denotes best per column. ↑ higher is better, ↓ lower is better.

# Model Pearson r ↑ Spearman ρ ↑ Kendall τ ↑ R² ↑ RMSE ↓ MAE ↓ Mean Error
1 DeepLC 0.804 ± 0.013 0.849 ± 0.008 0.671 ± 0.011 0.579 ± 0.033 6.692 ± 0.263 4.644 ± 0.214 −2.477 ± 0.592
2 GIN 0.788 ± 0.014 0.838 ± 0.015 0.661 ± 0.016 0.584 ± 0.038 6.649 ± 0.300 4.474 ± 0.177 −1.798 ± 0.634
3 ESM3-small 0.787 ± 0.008 0.826 ± 0.007 0.638 ± 0.009 0.399 ± 0.080 7.981 ± 0.539 5.622 ± 0.439 −3.996 ± 1.071
4 ESMC-300M 0.770 ± 0.017 0.828 ± 0.017 0.644 ± 0.019 0.406 ± 0.034 7.948 ± 0.227 5.468 ± 0.195 −4.088 ± 0.301
5 DeepRT-CapsNet 0.767 ± 0.022 0.804 ± 0.020 0.613 ± 0.023 0.521 ± 0.038 7.140 ± 0.283 5.103 ± 0.207 −2.028 ± 0.807
6 ESMC-600M 0.762 ± 0.012 0.826 ± 0.009 0.642 ± 0.010 0.321 ± 0.032 8.501 ± 0.198 6.023 ± 0.198 −4.942 ± 0.366
7 Transformer 0.742 ± 0.033 0.794 ± 0.028 0.605 ± 0.031 0.513 ± 0.053 7.188 ± 0.395 5.105 ± 0.320 −1.617 ± 0.901
8 Pretrained GIN 0.725 ± 0.015 0.778 ± 0.016 0.591 ± 0.016 0.460 ± 0.027 7.582 ± 0.189 5.234 ± 0.150 −2.619 ± 0.236
9 Morgan FP MLP 0.506 ± 0.021 0.534 ± 0.026 0.377 ± 0.019 0.173 ± 0.032 9.384 ± 0.181 6.866 ± 0.155 −1.745 ± 0.409

Diastereomer Pair Performance — D/L-Phe Pairs

Mean ± std over seeds on the test set. Bold denotes best per column. ↑ higher is better, ↓ lower is better.

# Model Pairwise Acc. ↑ Δ Pearson ↑ Δ Spearman ↑ Δ Kendall ↑ Δ AUC ↑ Δ RMSE ↓ Δ MAE ↓
1 GIN 0.665 ± 0.035 0.107 ± 0.145 0.136 ± 0.156 0.103 ± 0.108 0.539 ± 0.095 7.238 ± 0.430 5.208 ± 0.391
2 Pretrained GIN 0.646 ± 0.013 0.087 ± 0.041 0.134 ± 0.061 0.095 ± 0.045 0.580 ± 0.043 7.138 ± 0.061 5.119 ± 0.078
3 DeepLC 0.624 ± 0.083 0.009 ± 0.107 0.026 ± 0.128 0.029 ± 0.090 0.471 ± 0.076 7.243 ± 0.231 5.331 ± 0.269
4 Morgan FP MLP 0.619 ± 0.023 −0.038 ± 0.050 −0.028 ± 0.042 −0.023 ± 0.029 0.524 ± 0.024 7.789 ± 0.186 5.639 ± 0.123
5 Transformer 0.602 ± 0.062 0.100 ± 0.106 0.099 ± 0.127 0.065 ± 0.085 0.556 ± 0.090 7.360 ± 0.452 5.406 ± 0.428
6 ESM3-small 0.600 ± 0.040 0.061 ± 0.070 0.063 ± 0.084 0.042 ± 0.055 0.515 ± 0.048 7.319 ± 0.078 5.512 ± 0.086
7 ESMC-300M 0.600 ± 0.026 0.086 ± 0.045 0.125 ± 0.059 0.087 ± 0.040 0.547 ± 0.047 7.213 ± 0.069 5.327 ± 0.095
8 ESMC-600M 0.563 ± 0.024 −0.029 ± 0.041 −0.029 ± 0.038 −0.020 ± 0.025 0.451 ± 0.023 7.342 ± 0.060 5.491 ± 0.050
9 DeepRT-CapsNet 0.434 ± 0.063 −0.071 ± 0.065 −0.068 ± 0.052 −0.047 ± 0.038 0.441 ± 0.025 7.894 ± 0.386 6.017 ± 0.383