Configuration

ESM-2-650M (DeepProtein benchmark)

ESM-2-650M: protein language model; default DeepProtein configuration.

Overview

An explanatory profile has not yet been reviewed for this record.

Consult the linked sources for architecture or protocol details. Missing evidence is not evidence of a missing capability.

Evaluations and results

26 evaluations · 26 results. Different protocols are not a single leaderboard.

Sorted by MAE (mean absolute error) (lower is better). The best value in each column is highlighted. Decimals are rounded for display; each value links to the printed value and its source.

Tested configuration	Protocol and dataset	MAE (mean absolute error)	Macro F1 (averaged macro F1 (Figure 3 caption: 'averaged macro F1'), printed on a 0 to 100 scale)	AUPRC (area under the precision-recall curve, printed on a 0 to 100 scale)	R² (coefficient of determination)	Spearman (Spearman rank correlation, as defined in the article's Evaluation metrics section; supervised split)	Accuracy (accuracy, printed on a 0 to 100 scale)
ESM-2-650M (DeepProtein benchmark)Evaluation	DeepProtein CRISPR-Leenay repair outcome: MAE ↓, CRISPR-Leenay repair outcome	0.1218 ± 0.0010 (best in column)	Not reported	Not reported	Not reported	Not reported	Not reported
ESM-2-650M (DeepProtein benchmark)Evaluation	DeepProtein beta-lactamase (PEER): MAE ↓, beta-lactamase (PEER)	0.154 ± 0.001	Not reported	Not reported	Not reported	Not reported	Not reported
ESM-2-650M (DeepProtein benchmark)Evaluation	DeepProtein SAbDab-Chen antibody developability: MAE ↓, SAbDab-Chen antibody developability	0.224 ± 0.001	Not reported	Not reported	Not reported	Not reported	Not reported
ESM-2-650M (DeepProtein benchmark)Evaluation	DeepProtein Stability (TAPE): MAE ↓, Stability (TAPE)	0.273 ± 0.014	Not reported	Not reported	Not reported	Not reported	Not reported
ESM-2-650M (DeepProtein benchmark)Evaluation	DeepProtein Fluorescence (TAPE): MAE ↓, Fluorescence (TAPE)	0.563 ± 0.002	Not reported	Not reported	Not reported	Not reported	Not reported
ESM-2-650M (DeepProtein benchmark)Evaluation	DeepProtein PPI affinity (PEER): MAE ↓, PPI affinity (PEER)	1.893 ± 0.005	Not reported	Not reported	Not reported	Not reported	Not reported
ESM-2-650M (DeepProtein benchmark)Evaluation	DeepProtein TAP antibody developability: MAE ↓, TAP antibody developability	2.452 ± 0.027	Not reported	Not reported	Not reported	Not reported	Not reported
ESM-2-650M (DeepProtein benchmark)Evaluation	DeepProtein beta-lactamase (PEER): ρ ↑, beta-lactamase (PEER)	Not reported	Not reported	Not reported	Not reported	0.694 ± 0.001 (best in column)	Not reported
ESM-2-650M (DeepProtein benchmark)Evaluation	DeepProtein Binary localization (PEER): Averaged F1 ↑, Binary localization (PEER)	Not reported	91.76 ± 0.47% (best in column)	Not reported	Not reported	Not reported	Not reported
ESM-2-650M (DeepProtein benchmark)Evaluation	DeepProtein Binary localization (PEER): PR-AUC ↑, Binary localization (PEER)	Not reported	Not reported	96.63 ± 0.18% (best in column)	Not reported	Not reported	Not reported
ESM-2-650M (DeepProtein benchmark)Evaluation	DeepProtein CRISPR-Leenay repair outcome: R2 ↑, CRISPR-Leenay repair outcome	Not reported	Not reported	Not reported	0.110 ± 0.011	Not reported	Not reported
ESM-2-650M (DeepProtein benchmark)Evaluation	DeepProtein Fluorescence (TAPE): ρ ↑, Fluorescence (TAPE)	Not reported	Not reported	Not reported	Not reported	0.548 ± 0.004	Not reported
ESM-2-650M (DeepProtein benchmark)Evaluation	DeepProtein Fold classification (TAPE): Acc ↑, Fold classification (TAPE)	Not reported	Not reported	Not reported	Not reported	Not reported	49.80 ± 0.36%
ESM-2-650M (DeepProtein benchmark)Evaluation	DeepProtein Fold classification (TAPE): Averaged F1 ↑, Fold classification (TAPE)	Not reported	25.55 ± 0.61%	Not reported	Not reported	Not reported	Not reported
ESM-2-650M (DeepProtein benchmark)Evaluation	DeepProtein Human PPI (PEER): Averaged F1 ↑, Human PPI (PEER)	Not reported	87.63 ± 0.60%	Not reported	Not reported	Not reported	Not reported
ESM-2-650M (DeepProtein benchmark)Evaluation	DeepProtein Human PPI (PEER): PR-AUC ↑, Human PPI (PEER)	Not reported	Not reported	96.17 ± 0.18%	Not reported	Not reported	Not reported
ESM-2-650M (DeepProtein benchmark)Evaluation	DeepProtein PPI affinity (PEER): R2 ↑, PPI affinity (PEER)	Not reported	Not reported	Not reported	0.592 ± 0.001	Not reported	Not reported
ESM-2-650M (DeepProtein benchmark)Evaluation	DeepProtein SAbDab-Chen antibody developability: R2 ↑, SAbDab-Chen antibody developability	Not reported	Not reported	Not reported	0.600 ± 0.001	Not reported	Not reported
ESM-2-650M (DeepProtein benchmark)Evaluation	DeepProtein Solubility (PEER): Averaged F1 ↑, Solubility (PEER)	Not reported	63.33 ± 3.68%	Not reported	Not reported	Not reported	Not reported
ESM-2-650M (DeepProtein benchmark)Evaluation	DeepProtein Solubility (PEER): PR-AUC ↑, Solubility (PEER)	Not reported	Not reported	71.53 ± 0.58%	Not reported	Not reported	Not reported
ESM-2-650M (DeepProtein benchmark)Evaluation	DeepProtein Stability (TAPE): ρ ↑, Stability (TAPE)	Not reported	Not reported	Not reported	Not reported	0.689 ± 0.020	Not reported
ESM-2-650M (DeepProtein benchmark)Evaluation	DeepProtein Subcellular localization (PEER): Acc ↑, Subcellular localization (PEER)	Not reported	Not reported	Not reported	Not reported	Not reported	79.07 ± 0.05% (best in column)
ESM-2-650M (DeepProtein benchmark)Evaluation	DeepProtein Subcellular localization (PEER): Averaged F1 ↑, Subcellular localization (PEER)	Not reported	66.66 ± 1.14%	Not reported	Not reported	Not reported	Not reported
ESM-2-650M (DeepProtein benchmark)Evaluation	DeepProtein TAP antibody developability: R2 ↑, TAP antibody developability	Not reported	Not reported	Not reported	0.866 ± 0.007 (best in column)	Not reported	Not reported
ESM-2-650M (DeepProtein benchmark)Evaluation	DeepProtein Yeast PPI (PEER): Averaged F1 ↑, Yeast PPI (PEER)	Not reported	63.99 ± 1.00%	Not reported	Not reported	Not reported	Not reported
ESM-2-650M (DeepProtein benchmark)Evaluation	DeepProtein Yeast PPI (PEER): PR-AUC ↑, Yeast PPI (PEER)	Not reported	Not reported	67.36 ± 0.80%	Not reported	Not reported	Not reported

All 26 result rows with coverage, uncertainty and sources

Filter evaluations

Tested model or methodProtocolDatasetEvaluation setupMetricEvidence origin

Applied filters: All linked evaluations

Exact evaluated configurations and original reported results
Tested configuration	Protocol and dataset	Finding	Evidence and details
Configuration: ESM-2-650M (DeepProtein benchmark)	Protocol: DeepProtein beta-lactamase (PEER): MAE ↓ Dataset: beta-lactamase (PEER)	0.154 ± 0.001 MAE (mean absolute error) unit-unreported · lower Uncertainty: SD 0.001 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Figure 3's caption calls the regression metric 'Spearman (Pearson) Coefficient' while the Evaluation metrics section defines rho as Spearman rank correlation and R2 separately. The section's definition is taken as authoritative and the rho columns are recorded as Spearman; the caption wording is noted so a reader of the figure is not surprised.DeepProtein: deep learning library and benchmark for protein sequence learning · Tables 2, 4 and 6 column headings 'rho' and 'R2'; Figure 3 caption Methods, coverage and source ESM-2-650M on DeepProtein beta-lactamase (PEER) (MAE ↓) lit-protein-20261011-deepprotein-protocol-beta-lactamase-mae Aggregation: Not reported DeepProtein: deep learning library and benchmark for protein sequence learning · Table 2. row 'ESM-2-650M', column group 'β-lactamase', column 'MAE ↓'
Configuration: ESM-2-650M (DeepProtein benchmark)	Protocol: DeepProtein beta-lactamase (PEER): ρ ↑ Dataset: beta-lactamase (PEER)	0.694 ± 0.001 Spearman (Spearman rank correlation, as defined in the article's Evaluation metrics section; supervised split) unitless · higher Uncertainty: SD 0.001 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Figure 3's caption calls the regression metric 'Spearman (Pearson) Coefficient' while the Evaluation metrics section defines rho as Spearman rank correlation and R2 separately. The section's definition is taken as authoritative and the rho columns are recorded as Spearman; the caption wording is noted so a reader of the figure is not surprised.DeepProtein: deep learning library and benchmark for protein sequence learning · Tables 2, 4 and 6 column headings 'rho' and 'R2'; Figure 3 caption Methods, coverage and source ESM-2-650M on DeepProtein beta-lactamase (PEER) (ρ ↑) lit-protein-20261011-deepprotein-protocol-beta-lactamase-rho Aggregation: Not reported DeepProtein: deep learning library and benchmark for protein sequence learning · Table 2. row 'ESM-2-650M', column group 'β-lactamase', column 'ρ ↑'
Configuration: ESM-2-650M (DeepProtein benchmark)	Protocol: DeepProtein Binary localization (PEER): Averaged F1 ↑ Dataset: Binary localization (PEER)	91.76 ± 0.47% Macro F1 (averaged macro F1 (Figure 3 caption: 'averaged macro F1'), printed on a 0 to 100 scale) percent · higher Uncertainty: SD 0.47 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Figure 3's caption calls the regression metric 'Spearman (Pearson) Coefficient' while the Evaluation metrics section defines rho as Spearman rank correlation and R2 separately. The section's definition is taken as authoritative and the rho columns are recorded as Spearman; the caption wording is noted so a reader of the figure is not surprised.DeepProtein: deep learning library and benchmark for protein sequence learning · Tables 2, 4 and 6 column headings 'rho' and 'R2'; Figure 3 caption Methods, coverage and source ESM-2-650M on DeepProtein Binary localization (PEER) (Averaged F1 ↑) lit-protein-20261011-deepprotein-protocol-binary-avg-f1 Aggregation: Not reported DeepProtein: deep learning library and benchmark for protein sequence learning · Table 3. row 'ESM-2-650M', column group 'Binary', column 'Averaged F1 ↑'
Configuration: ESM-2-650M (DeepProtein benchmark)	Protocol: DeepProtein Binary localization (PEER): PR-AUC ↑ Dataset: Binary localization (PEER)	96.63 ± 0.18% AUPRC (area under the precision-recall curve, printed on a 0 to 100 scale) percent · higher Uncertainty: SD 0.18 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Figure 3's caption calls the regression metric 'Spearman (Pearson) Coefficient' while the Evaluation metrics section defines rho as Spearman rank correlation and R2 separately. The section's definition is taken as authoritative and the rho columns are recorded as Spearman; the caption wording is noted so a reader of the figure is not surprised.DeepProtein: deep learning library and benchmark for protein sequence learning · Tables 2, 4 and 6 column headings 'rho' and 'R2'; Figure 3 caption Methods, coverage and source ESM-2-650M on DeepProtein Binary localization (PEER) (PR-AUC ↑) lit-protein-20261011-deepprotein-protocol-binary-pr-auc Aggregation: Not reported DeepProtein: deep learning library and benchmark for protein sequence learning · Table 3. row 'ESM-2-650M', column group 'Binary', column 'PR-AUC ↑'
Configuration: ESM-2-650M (DeepProtein benchmark)	Protocol: DeepProtein CRISPR-Leenay repair outcome: MAE ↓ Dataset: CRISPR-Leenay repair outcome	0.1218 ± 0.0010 MAE (mean absolute error) unit-unreported · lower Uncertainty: SD 0.0010 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Figure 3's caption calls the regression metric 'Spearman (Pearson) Coefficient' while the Evaluation metrics section defines rho as Spearman rank correlation and R2 separately. The section's definition is taken as authoritative and the rho columns are recorded as Spearman; the caption wording is noted so a reader of the figure is not surprised.DeepProtein: deep learning library and benchmark for protein sequence learning · Tables 2, 4 and 6 column headings 'rho' and 'R2'; Figure 3 caption Methods, coverage and source ESM-2-650M on DeepProtein CRISPR-Leenay repair outcome (MAE ↓) lit-protein-20261011-deepprotein-protocol-crispr-leenay-mae Aggregation: Not reported DeepProtein: deep learning library and benchmark for protein sequence learning · Table 6. row 'ESM-2-650M', column group 'CRISPR-Leenay', column 'MAE ↓'
Configuration: ESM-2-650M (DeepProtein benchmark)	Protocol: DeepProtein CRISPR-Leenay repair outcome: R2 ↑ Dataset: CRISPR-Leenay repair outcome	0.110 ± 0.011 R² (coefficient of determination) unitless · higher Uncertainty: SD 0.011 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Figure 3's caption calls the regression metric 'Spearman (Pearson) Coefficient' while the Evaluation metrics section defines rho as Spearman rank correlation and R2 separately. The section's definition is taken as authoritative and the rho columns are recorded as Spearman; the caption wording is noted so a reader of the figure is not surprised.DeepProtein: deep learning library and benchmark for protein sequence learning · Tables 2, 4 and 6 column headings 'rho' and 'R2'; Figure 3 caption Methods, coverage and source ESM-2-650M on DeepProtein CRISPR-Leenay repair outcome (R2 ↑) lit-protein-20261011-deepprotein-protocol-crispr-leenay-r2 Aggregation: Not reported DeepProtein: deep learning library and benchmark for protein sequence learning · Table 6. row 'ESM-2-650M', column group 'CRISPR-Leenay', column 'R2 ↑'
Configuration: ESM-2-650M (DeepProtein benchmark)	Protocol: DeepProtein Fluorescence (TAPE): MAE ↓ Dataset: Fluorescence (TAPE)	0.563 ± 0.002 MAE (mean absolute error) unit-unreported · lower Uncertainty: SD 0.002 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Figure 3's caption calls the regression metric 'Spearman (Pearson) Coefficient' while the Evaluation metrics section defines rho as Spearman rank correlation and R2 separately. The section's definition is taken as authoritative and the rho columns are recorded as Spearman; the caption wording is noted so a reader of the figure is not surprised.DeepProtein: deep learning library and benchmark for protein sequence learning · Tables 2, 4 and 6 column headings 'rho' and 'R2'; Figure 3 caption Methods, coverage and source ESM-2-650M on DeepProtein Fluorescence (TAPE) (MAE ↓) lit-protein-20261011-deepprotein-protocol-fluorescence-mae Aggregation: Not reported DeepProtein: deep learning library and benchmark for protein sequence learning · Table 2. row 'ESM-2-650M', column group 'Fluorescence', column 'MAE ↓'
Configuration: ESM-2-650M (DeepProtein benchmark)	Protocol: DeepProtein Fluorescence (TAPE): ρ ↑ Dataset: Fluorescence (TAPE)	0.548 ± 0.004 Spearman (Spearman rank correlation, as defined in the article's Evaluation metrics section; supervised split) unitless · higher Uncertainty: SD 0.004 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Figure 3's caption calls the regression metric 'Spearman (Pearson) Coefficient' while the Evaluation metrics section defines rho as Spearman rank correlation and R2 separately. The section's definition is taken as authoritative and the rho columns are recorded as Spearman; the caption wording is noted so a reader of the figure is not surprised.DeepProtein: deep learning library and benchmark for protein sequence learning · Tables 2, 4 and 6 column headings 'rho' and 'R2'; Figure 3 caption Methods, coverage and source ESM-2-650M on DeepProtein Fluorescence (TAPE) (ρ ↑) lit-protein-20261011-deepprotein-protocol-fluorescence-rho Aggregation: Not reported DeepProtein: deep learning library and benchmark for protein sequence learning · Table 2. row 'ESM-2-650M', column group 'Fluorescence', column 'ρ ↑'
Configuration: ESM-2-650M (DeepProtein benchmark)	Protocol: DeepProtein Fold classification (TAPE): Acc ↑ Dataset: Fold classification (TAPE)	49.80 ± 0.36% Accuracy (accuracy, printed on a 0 to 100 scale) percent · higher Uncertainty: SD 0.36 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Figure 3's caption calls the regression metric 'Spearman (Pearson) Coefficient' while the Evaluation metrics section defines rho as Spearman rank correlation and R2 separately. The section's definition is taken as authoritative and the rho columns are recorded as Spearman; the caption wording is noted so a reader of the figure is not surprised.DeepProtein: deep learning library and benchmark for protein sequence learning · Tables 2, 4 and 6 column headings 'rho' and 'R2'; Figure 3 caption Methods, coverage and source ESM-2-650M on DeepProtein Fold classification (TAPE) (Acc ↑) lit-protein-20261011-deepprotein-protocol-fold-acc Aggregation: Not reported DeepProtein: deep learning library and benchmark for protein sequence learning · Table 7. row 'ESM-2-650M', column group 'Fold', column 'Acc ↑'
Configuration: ESM-2-650M (DeepProtein benchmark)	Protocol: DeepProtein Fold classification (TAPE): Averaged F1 ↑ Dataset: Fold classification (TAPE)	25.55 ± 0.61% Macro F1 (averaged macro F1 (Figure 3 caption: 'averaged macro F1'), printed on a 0 to 100 scale) percent · higher Uncertainty: SD 0.61 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Figure 3's caption calls the regression metric 'Spearman (Pearson) Coefficient' while the Evaluation metrics section defines rho as Spearman rank correlation and R2 separately. The section's definition is taken as authoritative and the rho columns are recorded as Spearman; the caption wording is noted so a reader of the figure is not surprised.DeepProtein: deep learning library and benchmark for protein sequence learning · Tables 2, 4 and 6 column headings 'rho' and 'R2'; Figure 3 caption Methods, coverage and source ESM-2-650M on DeepProtein Fold classification (TAPE) (Averaged F1 ↑) lit-protein-20261011-deepprotein-protocol-fold-avg-f1 Aggregation: Not reported DeepProtein: deep learning library and benchmark for protein sequence learning · Table 7. row 'ESM-2-650M', column group 'Fold', column 'Averaged F1 ↑'
Configuration: ESM-2-650M (DeepProtein benchmark)	Protocol: DeepProtein Human PPI (PEER): Averaged F1 ↑ Dataset: Human PPI (PEER)	87.63 ± 0.60% Macro F1 (averaged macro F1 (Figure 3 caption: 'averaged macro F1'), printed on a 0 to 100 scale) percent · higher Uncertainty: SD 0.60 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Figure 3's caption calls the regression metric 'Spearman (Pearson) Coefficient' while the Evaluation metrics section defines rho as Spearman rank correlation and R2 separately. The section's definition is taken as authoritative and the rho columns are recorded as Spearman; the caption wording is noted so a reader of the figure is not surprised.DeepProtein: deep learning library and benchmark for protein sequence learning · Tables 2, 4 and 6 column headings 'rho' and 'R2'; Figure 3 caption Methods, coverage and source ESM-2-650M on DeepProtein Human PPI (PEER) (Averaged F1 ↑) lit-protein-20261011-deepprotein-protocol-human-ppi-avg-f1 Aggregation: Not reported DeepProtein: deep learning library and benchmark for protein sequence learning · Table 4. row 'ESM-2-650M', column group 'Human PPI', column 'Averaged F1 ↑'
Configuration: ESM-2-650M (DeepProtein benchmark)	Protocol: DeepProtein Human PPI (PEER): PR-AUC ↑ Dataset: Human PPI (PEER)	96.17 ± 0.18% AUPRC (area under the precision-recall curve, printed on a 0 to 100 scale) percent · higher Uncertainty: SD 0.18 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Figure 3's caption calls the regression metric 'Spearman (Pearson) Coefficient' while the Evaluation metrics section defines rho as Spearman rank correlation and R2 separately. The section's definition is taken as authoritative and the rho columns are recorded as Spearman; the caption wording is noted so a reader of the figure is not surprised.DeepProtein: deep learning library and benchmark for protein sequence learning · Tables 2, 4 and 6 column headings 'rho' and 'R2'; Figure 3 caption Methods, coverage and source ESM-2-650M on DeepProtein Human PPI (PEER) (PR-AUC ↑) lit-protein-20261011-deepprotein-protocol-human-ppi-pr-auc Aggregation: Not reported DeepProtein: deep learning library and benchmark for protein sequence learning · Table 4. row 'ESM-2-650M', column group 'Human PPI', column 'PR-AUC ↑'
Configuration: ESM-2-650M (DeepProtein benchmark)	Protocol: DeepProtein PPI affinity (PEER): MAE ↓ Dataset: PPI affinity (PEER)	1.893 ± 0.005 MAE (mean absolute error) unit-unreported · lower Uncertainty: SD 0.005 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Figure 3's caption calls the regression metric 'Spearman (Pearson) Coefficient' while the Evaluation metrics section defines rho as Spearman rank correlation and R2 separately. The section's definition is taken as authoritative and the rho columns are recorded as Spearman; the caption wording is noted so a reader of the figure is not surprised.DeepProtein: deep learning library and benchmark for protein sequence learning · Tables 2, 4 and 6 column headings 'rho' and 'R2'; Figure 3 caption Methods, coverage and source ESM-2-650M on DeepProtein PPI affinity (PEER) (MAE ↓) lit-protein-20261011-deepprotein-protocol-ppi-affinity-mae Aggregation: Not reported DeepProtein: deep learning library and benchmark for protein sequence learning · Table 4. row 'ESM-2-650M', column group 'PPI Affinity', column 'MAE ↓'
Configuration: ESM-2-650M (DeepProtein benchmark)	Protocol: DeepProtein PPI affinity (PEER): R2 ↑ Dataset: PPI affinity (PEER)	0.592 ± 0.001 R² (coefficient of determination) unitless · higher Uncertainty: SD 0.001 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Figure 3's caption calls the regression metric 'Spearman (Pearson) Coefficient' while the Evaluation metrics section defines rho as Spearman rank correlation and R2 separately. The section's definition is taken as authoritative and the rho columns are recorded as Spearman; the caption wording is noted so a reader of the figure is not surprised.DeepProtein: deep learning library and benchmark for protein sequence learning · Tables 2, 4 and 6 column headings 'rho' and 'R2'; Figure 3 caption Methods, coverage and source ESM-2-650M on DeepProtein PPI affinity (PEER) (R2 ↑) lit-protein-20261011-deepprotein-protocol-ppi-affinity-r2 Aggregation: Not reported DeepProtein: deep learning library and benchmark for protein sequence learning · Table 4. row 'ESM-2-650M', column group 'PPI Affinity', column 'R2 ↑'
Configuration: ESM-2-650M (DeepProtein benchmark)	Protocol: DeepProtein SAbDab-Chen antibody developability: MAE ↓ Dataset: SAbDab-Chen antibody developability	0.224 ± 0.001 MAE (mean absolute error) unit-unreported · lower Uncertainty: SD 0.001 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Figure 3's caption calls the regression metric 'Spearman (Pearson) Coefficient' while the Evaluation metrics section defines rho as Spearman rank correlation and R2 separately. The section's definition is taken as authoritative and the rho columns are recorded as Spearman; the caption wording is noted so a reader of the figure is not surprised.DeepProtein: deep learning library and benchmark for protein sequence learning · Tables 2, 4 and 6 column headings 'rho' and 'R2'; Figure 3 caption Methods, coverage and source ESM-2-650M on DeepProtein SAbDab-Chen antibody developability (MAE ↓) lit-protein-20261011-deepprotein-protocol-sabdab-chen-mae Aggregation: Not reported DeepProtein: deep learning library and benchmark for protein sequence learning · Table 6. row 'ESM-2-650M', column group 'SAbDab-Chen', column 'MAE ↓'
Configuration: ESM-2-650M (DeepProtein benchmark)	Protocol: DeepProtein SAbDab-Chen antibody developability: R2 ↑ Dataset: SAbDab-Chen antibody developability	0.600 ± 0.001 R² (coefficient of determination) unitless · higher Uncertainty: SD 0.001 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Figure 3's caption calls the regression metric 'Spearman (Pearson) Coefficient' while the Evaluation metrics section defines rho as Spearman rank correlation and R2 separately. The section's definition is taken as authoritative and the rho columns are recorded as Spearman; the caption wording is noted so a reader of the figure is not surprised.DeepProtein: deep learning library and benchmark for protein sequence learning · Tables 2, 4 and 6 column headings 'rho' and 'R2'; Figure 3 caption Methods, coverage and source ESM-2-650M on DeepProtein SAbDab-Chen antibody developability (R2 ↑) lit-protein-20261011-deepprotein-protocol-sabdab-chen-r2 Aggregation: Not reported DeepProtein: deep learning library and benchmark for protein sequence learning · Table 6. row 'ESM-2-650M', column group 'SAbDab-Chen', column 'R2 ↑'
Configuration: ESM-2-650M (DeepProtein benchmark)	Protocol: DeepProtein Solubility (PEER): Averaged F1 ↑ Dataset: Solubility (PEER)	63.33 ± 3.68% Macro F1 (averaged macro F1 (Figure 3 caption: 'averaged macro F1'), printed on a 0 to 100 scale) percent · higher Uncertainty: SD 3.68 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Figure 3's caption calls the regression metric 'Spearman (Pearson) Coefficient' while the Evaluation metrics section defines rho as Spearman rank correlation and R2 separately. The section's definition is taken as authoritative and the rho columns are recorded as Spearman; the caption wording is noted so a reader of the figure is not surprised.DeepProtein: deep learning library and benchmark for protein sequence learning · Tables 2, 4 and 6 column headings 'rho' and 'R2'; Figure 3 caption Methods, coverage and source ESM-2-650M on DeepProtein Solubility (PEER) (Averaged F1 ↑) lit-protein-20261011-deepprotein-protocol-solubility-avg-f1 Aggregation: Not reported DeepProtein: deep learning library and benchmark for protein sequence learning · Table 2. row 'ESM-2-650M', column group 'Solubility', column 'Averaged F1 ↑'
Configuration: ESM-2-650M (DeepProtein benchmark)	Protocol: DeepProtein Solubility (PEER): PR-AUC ↑ Dataset: Solubility (PEER)	71.53 ± 0.58% AUPRC (area under the precision-recall curve, printed on a 0 to 100 scale) percent · higher Uncertainty: SD 0.58 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Figure 3's caption calls the regression metric 'Spearman (Pearson) Coefficient' while the Evaluation metrics section defines rho as Spearman rank correlation and R2 separately. The section's definition is taken as authoritative and the rho columns are recorded as Spearman; the caption wording is noted so a reader of the figure is not surprised.DeepProtein: deep learning library and benchmark for protein sequence learning · Tables 2, 4 and 6 column headings 'rho' and 'R2'; Figure 3 caption Methods, coverage and source ESM-2-650M on DeepProtein Solubility (PEER) (PR-AUC ↑) lit-protein-20261011-deepprotein-protocol-solubility-pr-auc Aggregation: Not reported DeepProtein: deep learning library and benchmark for protein sequence learning · Table 2. row 'ESM-2-650M', column group 'Solubility', column 'PR-AUC ↑'
Configuration: ESM-2-650M (DeepProtein benchmark)	Protocol: DeepProtein Stability (TAPE): MAE ↓ Dataset: Stability (TAPE)	0.273 ± 0.014 MAE (mean absolute error) unit-unreported · lower Uncertainty: SD 0.014 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Figure 3's caption calls the regression metric 'Spearman (Pearson) Coefficient' while the Evaluation metrics section defines rho as Spearman rank correlation and R2 separately. The section's definition is taken as authoritative and the rho columns are recorded as Spearman; the caption wording is noted so a reader of the figure is not surprised.DeepProtein: deep learning library and benchmark for protein sequence learning · Tables 2, 4 and 6 column headings 'rho' and 'R2'; Figure 3 caption Methods, coverage and source ESM-2-650M on DeepProtein Stability (TAPE) (MAE ↓) lit-protein-20261011-deepprotein-protocol-stability-mae Aggregation: Not reported DeepProtein: deep learning library and benchmark for protein sequence learning · Table 2. row 'ESM-2-650M', column group 'Stability', column 'MAE ↓'
Configuration: ESM-2-650M (DeepProtein benchmark)	Protocol: DeepProtein Stability (TAPE): ρ ↑ Dataset: Stability (TAPE)	0.689 ± 0.020 Spearman (Spearman rank correlation, as defined in the article's Evaluation metrics section; supervised split) unitless · higher Uncertainty: SD 0.020 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Figure 3's caption calls the regression metric 'Spearman (Pearson) Coefficient' while the Evaluation metrics section defines rho as Spearman rank correlation and R2 separately. The section's definition is taken as authoritative and the rho columns are recorded as Spearman; the caption wording is noted so a reader of the figure is not surprised.DeepProtein: deep learning library and benchmark for protein sequence learning · Tables 2, 4 and 6 column headings 'rho' and 'R2'; Figure 3 caption Methods, coverage and source ESM-2-650M on DeepProtein Stability (TAPE) (ρ ↑) lit-protein-20261011-deepprotein-protocol-stability-rho Aggregation: Not reported DeepProtein: deep learning library and benchmark for protein sequence learning · Table 2. row 'ESM-2-650M', column group 'Stability', column 'ρ ↑'
Configuration: ESM-2-650M (DeepProtein benchmark)	Protocol: DeepProtein Subcellular localization (PEER): Acc ↑ Dataset: Subcellular localization (PEER)	79.07 ± 0.05% Accuracy (accuracy, printed on a 0 to 100 scale) percent · higher Uncertainty: SD 0.05 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Figure 3's caption calls the regression metric 'Spearman (Pearson) Coefficient' while the Evaluation metrics section defines rho as Spearman rank correlation and R2 separately. The section's definition is taken as authoritative and the rho columns are recorded as Spearman; the caption wording is noted so a reader of the figure is not surprised.DeepProtein: deep learning library and benchmark for protein sequence learning · Tables 2, 4 and 6 column headings 'rho' and 'R2'; Figure 3 caption Methods, coverage and source ESM-2-650M on DeepProtein Subcellular localization (PEER) (Acc ↑) lit-protein-20261011-deepprotein-protocol-subcellular-acc Aggregation: Not reported DeepProtein: deep learning library and benchmark for protein sequence learning · Table 3. row 'ESM-2-650M', column group 'Subcellular', column 'Acc ↑'
Configuration: ESM-2-650M (DeepProtein benchmark)	Protocol: DeepProtein Subcellular localization (PEER): Averaged F1 ↑ Dataset: Subcellular localization (PEER)	66.66 ± 1.14% Macro F1 (averaged macro F1 (Figure 3 caption: 'averaged macro F1'), printed on a 0 to 100 scale) percent · higher Uncertainty: SD 1.14 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Figure 3's caption calls the regression metric 'Spearman (Pearson) Coefficient' while the Evaluation metrics section defines rho as Spearman rank correlation and R2 separately. The section's definition is taken as authoritative and the rho columns are recorded as Spearman; the caption wording is noted so a reader of the figure is not surprised.DeepProtein: deep learning library and benchmark for protein sequence learning · Tables 2, 4 and 6 column headings 'rho' and 'R2'; Figure 3 caption Methods, coverage and source ESM-2-650M on DeepProtein Subcellular localization (PEER) (Averaged F1 ↑) lit-protein-20261011-deepprotein-protocol-subcellular-avg-f1 Aggregation: Not reported DeepProtein: deep learning library and benchmark for protein sequence learning · Table 3. row 'ESM-2-650M', column group 'Subcellular', column 'Averaged F1 ↑'
Configuration: ESM-2-650M (DeepProtein benchmark)	Protocol: DeepProtein TAP antibody developability: MAE ↓ Dataset: TAP antibody developability	2.452 ± 0.027 MAE (mean absolute error) unit-unreported · lower Uncertainty: SD 0.027 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Figure 3's caption calls the regression metric 'Spearman (Pearson) Coefficient' while the Evaluation metrics section defines rho as Spearman rank correlation and R2 separately. The section's definition is taken as authoritative and the rho columns are recorded as Spearman; the caption wording is noted so a reader of the figure is not surprised.DeepProtein: deep learning library and benchmark for protein sequence learning · Tables 2, 4 and 6 column headings 'rho' and 'R2'; Figure 3 caption Methods, coverage and source ESM-2-650M on DeepProtein TAP antibody developability (MAE ↓) lit-protein-20261011-deepprotein-protocol-tap-mae Aggregation: Not reported DeepProtein: deep learning library and benchmark for protein sequence learning · Table 6. row 'ESM-2-650M', column group 'TAP', column 'MAE ↓'
Configuration: ESM-2-650M (DeepProtein benchmark)	Protocol: DeepProtein TAP antibody developability: R2 ↑ Dataset: TAP antibody developability	0.866 ± 0.007 R² (coefficient of determination) unitless · higher Uncertainty: SD 0.007 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Figure 3's caption calls the regression metric 'Spearman (Pearson) Coefficient' while the Evaluation metrics section defines rho as Spearman rank correlation and R2 separately. The section's definition is taken as authoritative and the rho columns are recorded as Spearman; the caption wording is noted so a reader of the figure is not surprised.DeepProtein: deep learning library and benchmark for protein sequence learning · Tables 2, 4 and 6 column headings 'rho' and 'R2'; Figure 3 caption Methods, coverage and source ESM-2-650M on DeepProtein TAP antibody developability (R2 ↑) lit-protein-20261011-deepprotein-protocol-tap-r2 Aggregation: Not reported DeepProtein: deep learning library and benchmark for protein sequence learning · Table 6. row 'ESM-2-650M', column group 'TAP', column 'R2 ↑'
Configuration: ESM-2-650M (DeepProtein benchmark)	Protocol: DeepProtein Yeast PPI (PEER): Averaged F1 ↑ Dataset: Yeast PPI (PEER)	63.99 ± 1.00% Macro F1 (averaged macro F1 (Figure 3 caption: 'averaged macro F1'), printed on a 0 to 100 scale) percent · higher Uncertainty: SD 1.00 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Figure 3's caption calls the regression metric 'Spearman (Pearson) Coefficient' while the Evaluation metrics section defines rho as Spearman rank correlation and R2 separately. The section's definition is taken as authoritative and the rho columns are recorded as Spearman; the caption wording is noted so a reader of the figure is not surprised.DeepProtein: deep learning library and benchmark for protein sequence learning · Tables 2, 4 and 6 column headings 'rho' and 'R2'; Figure 3 caption Methods, coverage and source ESM-2-650M on DeepProtein Yeast PPI (PEER) (Averaged F1 ↑) lit-protein-20261011-deepprotein-protocol-yeast-ppi-avg-f1 Aggregation: Not reported DeepProtein: deep learning library and benchmark for protein sequence learning · Table 4. row 'ESM-2-650M', column group 'Yeast PPI', column 'Averaged F1 ↑'

Source checking is not independent reproduction. Release 2026-10-22-1b7cf027e4f6.

Use this configuration

How it works, versions and access

Strengths, limitations and unresolved questions

Evidence

Source checking verifies the cited claim or transcription. It does not establish independent reproduction.

Evidence table

Inspect claims, sources and review details

Trace each statement to its source and review. A context-only reference supports the record generally; it does not verify an individual field. Source checking does not reproduce an experiment.

One row per statement and cited source. Multiple citations are not independent evaluations. Shared locators are labelled explicitly.

0 evidence rows matching the loaded filters

Claims, original sources and review scope · Release 2026-10-22-1b7cf027e4f6
Property and statement	Original source and location	Review and provenance

No evidence rows match these filters. Choose another scope or clear the search.

How to read the evidence

Sources and history

Release 2026-10-22-1b7cf027e4f6 · Record review: source checked

1 source record and release history

DeepProtein: deep learning library and benchmark for protein sequence learning · Original source · Version of record, Bioinformatics 2025 (PMC12553333 JATS full text)

Download this release (gzip)

Technical metadata and extraction receipts

Stable ID: lit-protein-20261011-deepprotein-config-esm-2-650m

areas: protein-function
contexts: research
method types: foundation_model
reported name: ESM-2-650M
reported configuration: protein language model; default DeepProtein configuration (Table 9)
foundation model eligible: true
source locator: Tables 2 to 8, row 'ESM-2-650M'; Table 9 for the default configuration
missing metadata: version: reason: unreported; note: The tables name the model without a release version
provenance note: ESM-2-650M as trained and evaluated in the DeepProtein benchmark (protein language model).

Related records

Suggest a correction