Models / protein-ml / esm2-t33-650m-ur50d / benchmarks

protein-ml/esm2-t33-650m-ur50d

Visibility: PUBLIC

Model card FilesBenchmarks

Group by

Domain variant fitness, human PLM 0.5 (ProteinGym x Pfam) protein-ml/protein-domain-variant-fitness-pfam-05

Drawn from: ProteinGym, ProteinGym_v1

atlas-n25-v2

Metric

Value

95% CI

Rank

Best in catalogue

Source

Overlap

Spearman ρ (zero-shot, domain input) ↑ higher is better · zero-shot

By assay: 4 levels

assay	Value [95% CI]	n
CCR5_HUMAN_Gill_2023	0.238 [−0.205, 0.639]	25
MET_HUMAN_Estevam_2023	0.256 [−0.155, 0.616]	25
S22A1_HUMAN_Yee_2023_activity	0.472 [0.103, 0.734]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.275 [−0.086, 0.560]	25

0.310

[0.093, 0.475]

100

7 / 13

0.476

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

AUC (zero-shot, domain input) ↑ higher is better · zero-shot

By assay: 4 levels

assay	Value [95% CI]	n
CCR5_HUMAN_Gill_2023	0.603 [0.365, 0.813]	25
MET_HUMAN_Estevam_2023	0.675 [0.452, 0.885]	25
S22A1_HUMAN_Yee_2023_activity	0.700 [0.472, 0.904]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.722 [0.500, 0.909]	25

0.675

[0.567, 0.773]

100

6 / 13

0.759

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

MCC (zero-shot, domain input) ↑ higher is better · zero-shot

By assay: 4 levels

assay	Value [95% CI]	n
CCR5_HUMAN_Gill_2023	0.199 [−0.294, 0.529]	25
MET_HUMAN_Estevam_2023	0.277 [−0.131, 0.686]	25
S22A1_HUMAN_Yee_2023_activity	0.294 [−0.140, 0.680]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.280 [−0.242, 0.714]	25

0.263

[0.017, 0.448]

100

7 / 13

0.426

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

NDCG, top 10 % (zero-shot, domain input) ↑ higher is better · zero-shot · secondary

By assay: 4 levels

assay	Value [95% CI]	n
CCR5_HUMAN_Gill_2023	0.300 [0.000, 0.751]	25
MET_HUMAN_Estevam_2023	0.563 [0.387, 0.988]	25
S22A1_HUMAN_Yee_2023_activity	0.898 [0.265, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.435 [0.070, 0.810]	25

0.549

[0.383, 0.722]

100

—

0.635

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

Top-10 % recall (zero-shot, domain input) ↑ higher is better · zero-shot · secondary

By assay: 4 levels

assay	Value [95% CI]	n
CCR5_HUMAN_Gill_2023	0.000 [0.000, 0.333]	25
MET_HUMAN_Estevam_2023	0.333 [0.000, 1.000]	25
S22A1_HUMAN_Yee_2023_activity	0.333 [0.000, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.000 [0.000, 0.667]	25

0.167 best

[0.000, 0.417]

100

—

0.167

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

Domain variant fitness, human PLM 0.9 (ProteinGym x Pfam) protein-ml/protein-domain-variant-fitness-pfam-09

Drawn from: ProteinGym, ProteinGym_v1

atlas-n25-v2

Metric

Value

95% CI

Rank

Best in catalogue

Source

Overlap

Spearman ρ (zero-shot, domain input) ↑ higher is better · zero-shot

By assay: 4 levels

assay	Value [95% CI]	n
CCR5_HUMAN_Gill_2023	0.238 [−0.205, 0.639]	25
MET_HUMAN_Estevam_2023	0.256 [−0.155, 0.616]	25
S22A1_HUMAN_Yee_2023_activity	0.472 [0.103, 0.734]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.275 [−0.086, 0.560]	25

0.310

[0.093, 0.475]

100

7 / 13

0.476

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

AUC (zero-shot, domain input) ↑ higher is better · zero-shot

By assay: 4 levels

assay	Value [95% CI]	n
CCR5_HUMAN_Gill_2023	0.603 [0.365, 0.813]	25
MET_HUMAN_Estevam_2023	0.675 [0.452, 0.885]	25
S22A1_HUMAN_Yee_2023_activity	0.700 [0.472, 0.904]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.722 [0.500, 0.909]	25

0.675

[0.567, 0.773]

100

6 / 13

0.759

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

MCC (zero-shot, domain input) ↑ higher is better · zero-shot

By assay: 4 levels

assay	Value [95% CI]	n
CCR5_HUMAN_Gill_2023	0.199 [−0.294, 0.529]	25
MET_HUMAN_Estevam_2023	0.277 [−0.131, 0.686]	25
S22A1_HUMAN_Yee_2023_activity	0.294 [−0.140, 0.680]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.280 [−0.242, 0.714]	25

0.263

[0.017, 0.448]

100

7 / 13

0.426

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

NDCG, top 10 % (zero-shot, domain input) ↑ higher is better · zero-shot · secondary

By assay: 4 levels

assay	Value [95% CI]	n
CCR5_HUMAN_Gill_2023	0.300 [0.000, 0.751]	25
MET_HUMAN_Estevam_2023	0.563 [0.387, 0.988]	25
S22A1_HUMAN_Yee_2023_activity	0.898 [0.265, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.435 [0.070, 0.810]	25

0.549

[0.383, 0.722]

100

—

0.635

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

Top-10 % recall (zero-shot, domain input) ↑ higher is better · zero-shot · secondary

By assay: 4 levels

assay	Value [95% CI]	n
CCR5_HUMAN_Gill_2023	0.000 [0.000, 0.333]	25
MET_HUMAN_Estevam_2023	0.333 [0.000, 1.000]	25
S22A1_HUMAN_Yee_2023_activity	0.333 [0.000, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.000 [0.000, 0.667]	25

0.167 best

[0.000, 0.417]

100

—

0.167

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

Domain variant fitness, human PLM Clans-0.5 (ProteinGym x Pfam) protein-ml/protein-domain-variant-fitness-clan-05

Drawn from: ProteinGym, ProteinGym_v1

atlas-n25-v2

Metric

Value

95% CI

Rank

Best in catalogue

Source

Overlap

Spearman ρ (zero-shot, domain input) ↑ higher is better · zero-shot

By assay: 4 levels

assay	Value [95% CI]	n
ADRB2_HUMAN_Jones_2020	0.078 [−0.337, 0.475]	25
CCR5_HUMAN_Gill_2023	0.238 [−0.205, 0.639]	25
OPSD_HUMAN_Wan_2019	0.433 [0.034, 0.758]	25
S22A1_HUMAN_Yee_2023_activity	0.472 [0.103, 0.734]	25

0.306

[0.101, 0.484]

100

8 / 13

0.463

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

AUC (zero-shot, domain input) ↑ higher is better · zero-shot

By assay: 4 levels

assay	Value [95% CI]	n
ADRB2_HUMAN_Jones_2020	0.545 [0.305, 0.786]	25
CCR5_HUMAN_Gill_2023	0.603 [0.365, 0.813]	25
OPSD_HUMAN_Wan_2019	0.750 [0.547, 0.931]	25
S22A1_HUMAN_Yee_2023_activity	0.700 [0.472, 0.904]	25

0.650

[0.528, 0.764]

100

8 / 13

0.751

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

MCC (zero-shot, domain input) ↑ higher is better · zero-shot

By assay: 4 levels

assay	Value [95% CI]	n
ADRB2_HUMAN_Jones_2020	0.116 [−0.353, 0.519]	25
CCR5_HUMAN_Gill_2023	0.199 [−0.294, 0.529]	25
OPSD_HUMAN_Wan_2019	0.447 [−0.054, 0.761]	25
S22A1_HUMAN_Yee_2023_activity	0.294 [−0.140, 0.680]	25

0.264

[0.009, 0.428]

100

3 / 13

0.426

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

NDCG, top 10 % (zero-shot, domain input) ↑ higher is better · zero-shot · secondary

By assay: 4 levels

assay	Value [95% CI]	n
ADRB2_HUMAN_Jones_2020	0.453 [0.266, 0.776]	25
CCR5_HUMAN_Gill_2023	0.300 [0.000, 0.751]	25
OPSD_HUMAN_Wan_2019	0.437 [0.007, 0.913]	25
S22A1_HUMAN_Yee_2023_activity	0.898 [0.265, 1.000]	25

0.522

[0.324, 0.728]

100

—

0.709

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

Top-10 % recall (zero-shot, domain input) ↑ higher is better · zero-shot · secondary

By assay: 4 levels

assay	Value [95% CI]	n
ADRB2_HUMAN_Jones_2020	0.000 [0.000, 0.600]	25
CCR5_HUMAN_Gill_2023	0.000 [0.000, 0.333]	25
OPSD_HUMAN_Wan_2019	0.000 [0.000, 0.801]	25
S22A1_HUMAN_Yee_2023_activity	0.333 [0.000, 1.000]	25

0.083

[0.000, 0.417]

100

—

0.250

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

Domain variant fitness, human PLM Clans-0.9 (ProteinGym x Pfam) protein-ml/protein-domain-variant-fitness-clan-09

Drawn from: ProteinGym, ProteinGym_v1

atlas-n25-v2

Metric

Value

95% CI

Rank

Best in catalogue

Source

Overlap

Spearman ρ (zero-shot, domain input) ↑ higher is better · zero-shot

By assay: 4 levels

assay	Value [95% CI]	n
MET_HUMAN_Estevam_2023	0.256 [−0.155, 0.616]	25
MK01_HUMAN_Brenan_2016	0.291 [−0.201, 0.668]	25
SCN5A_HUMAN_Glazer_2019	0.112 [−0.386, 0.531]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.275 [−0.086, 0.560]	25

0.233

[0.011, 0.422]

100

8 / 13

0.362

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

AUC (zero-shot, domain input) ↑ higher is better · zero-shot

By assay: 4 levels

assay	Value [95% CI]	n
MET_HUMAN_Estevam_2023	0.675 [0.452, 0.885]	25
MK01_HUMAN_Brenan_2016	0.787 [0.500, 0.985]	25
SCN5A_HUMAN_Glazer_2019	0.580 [0.333, 0.794]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.722 [0.500, 0.909]	25

0.691

[0.574, 0.803]

100

3 / 13

0.722

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

MCC (zero-shot, domain input) ↑ higher is better · zero-shot

By assay: 4 levels

assay	Value [95% CI]	n
MET_HUMAN_Estevam_2023	0.277 [−0.131, 0.686]	25
MK01_HUMAN_Brenan_2016	0.316 [−0.023, 0.721]	25
SCN5A_HUMAN_Glazer_2019	0.033 [−0.458, 0.447]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.280 [−0.242, 0.714]	25

0.227

[0.018, 0.456]

100

7 / 13

0.310

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

NDCG, top 10 % (zero-shot, domain input) ↑ higher is better · zero-shot · secondary

By assay: 4 levels

assay	Value [95% CI]	n
MET_HUMAN_Estevam_2023	0.563 [0.387, 0.988]	25
MK01_HUMAN_Brenan_2016	0.364 [0.000, 0.884]	25
SCN5A_HUMAN_Glazer_2019	0.710 [0.078, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.435 [0.070, 0.810]	25

0.518

[0.305, 0.769]

100

—

0.725

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

Top-10 % recall (zero-shot, domain input) ↑ higher is better · zero-shot · secondary

By assay: 4 levels

assay	Value [95% CI]	n
MET_HUMAN_Estevam_2023	0.333 [0.000, 1.000]	25
MK01_HUMAN_Brenan_2016	0.000 [0.000, 0.333]	25
SCN5A_HUMAN_Glazer_2019	0.667 [0.000, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.000 [0.000, 0.667]	25

0.250 best

[0.000, 0.458]

100

—

0.250

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

GO function prediction (Swiss-Prot 2025H2) protein-ml/protein-function-go-prediction

Drawn from: UniProt Knowledgebase

atlas-n100-v1

Metric	Value	95% CI	n	Rank	Best in catalogue	Source	Overlap	Board
Fmax (molecular function, probe) ↑ higher is better · probe	0.619	[0.562, 0.702]	100	16 / 17	0.844	reproduced	unknown	Board
Fmax (biological process, probe) ↑ higher is better · probe	0.436	[0.335, 0.550]	100	13 / 17	0.538	reproduced	unknown	Board
Fmax (cellular component, probe) ↑ higher is better · probe	0.754	[0.701, 0.814]	100	3 / 17	0.756	reproduced	unknown	Board
Micro AUPR (molecular function, probe) ↑ higher is better · probe · secondary	0.615	[0.532, 0.714]	100	—	0.875	reproduced	unknown	Board
Smin (molecular function, probe) ↓ lower is better · bits · probe · secondary	3.855	[2.998, 4.565]	100	—	1.727	reproduced	unknown	Board
Micro AUPR (biological process, probe) ↑ higher is better · probe · secondary	0.380	[0.240, 0.530]	100	—	0.595	reproduced	unknown	Board
Smin (biological process, probe) ↓ lower is better · bits · probe · secondary	5.861	[4.221, 7.756]	100	—	4.842	reproduced	unknown	Board
Micro AUPR (cellular component, probe) ↑ higher is better · probe · secondary	0.732	[0.668, 0.796]	100	—	0.758	reproduced	unknown	Board
Smin (cellular component, probe) ↓ lower is better · bits · probe · secondary	3.145	[2.440, 3.901]	100	—	3.092	reproduced	unknown	Board

Protein membrane association (DeepLoc) protein-ml/protein-membrane-localization

atlas-full-v1

Metric	Value	95% CI	n	Rank	Best in catalogue	Source	Overlap	Board
Q2 accuracy (logistic probe) ↑ higher is better · probe	0.849	[0.834, 0.867]	1728	6 / 17	0.888	reproduced	overlap Train/test overlap declared by the evaluator: protein-ml/uniref50	Board
MCC (logistic probe) ↑ higher is better · probe · secondary	0.690	[0.659, 0.727]	1728	—	0.771	reproduced	overlap Train/test overlap declared by the evaluator: protein-ml/uniref50	Board

Protein stability ΔΔG (Mega-scale) protein-ml/protein-stability-ddg

atlas-n100-v1

Metric	Value	95% CI	n	Rank	Best in catalogue	Source	Overlap	Board
Spearman ρ (zero-shot) ↑ higher is better · zero-shot	0.327	[0.056, 0.549]	100	5 / 12	0.410	reproduced	unknown	Board
Spearman ρ (ridge probe) ↑ higher is better · probe	0.253	[0.063, 0.423]	100	3 / 17	0.297	reproduced	unknown	Board
RMSE, kcal/mol (ridge probe) ↓ lower is better · kcal/mol · probe	1.277	[1.045, 1.485]	100	17 / 17	1.095	reproduced	unknown	Board
Pearson r (ridge probe) ↑ higher is better · probe · secondary	0.319	[0.103, 0.480]	100	—	0.349	reproduced	unknown	Board
MAE, kcal/mol (ridge probe) ↓ lower is better · kcal/mol · probe · secondary	1.017	[0.809, 1.217]	100	—	0.819	reproduced	unknown	Board

Protein subcellular localization (DeepLoc) protein-ml/protein-subcellular-localization

atlas-full-v1

Metric	Value	95% CI	n	Rank	Best in catalogue	Source	Overlap	Board
Accuracy (logistic probe) ↑ higher is better · probe	0.728	[0.713, 0.742]	2742	6 / 17	0.801	reproduced	overlap Train/test overlap declared by the evaluator: protein-ml/uniref50	Board
Macro F1 (logistic probe) ↑ higher is better · probe	0.545	[0.524, 0.564]	2742	4 / 17	0.640	reproduced	overlap Train/test overlap declared by the evaluator: protein-ml/uniref50	Board
MCC (logistic probe) ↑ higher is better · probe · secondary	0.669	[0.651, 0.687]	2742	—	0.758	reproduced	overlap Train/test overlap declared by the evaluator: protein-ml/uniref50	Board

ProteinGym DMS zero-shot (Spearman) protein-ml/proteingym-dms-zero-shot-spearman

Drawn from: ProteinGym, ProteinGym_v1

atlas-n100-v1

Metric

Value

95% CI

Rank

Best in catalogue

Source

Overlap

Spearman ρ (zero-shot) ↑ higher is better · zero-shot

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.573 [0.428, 0.717]	100
CASP3_HUMAN_Roychowdhury_2020	0.681 [0.573, 0.762]	100
CCR5_HUMAN_Gill_2023	0.311 [0.115, 0.511]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.639 [0.489, 0.740]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.263 [0.078, 0.450]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.096 [−0.099, 0.283]	100
MLAC_ECOLI_MacRae_2023	0.338 [0.173, 0.485]	100
MTH3_HAEAE_RockahShmuel_2015	0.555 [0.390, 0.683]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.663 [0.519, 0.778]	100
SPG1_STRSG_Olson_2014	0.130 [−0.083, 0.342]	100

0.425

[0.366, 0.472]

1000

2 / 11

0.434

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

AUC (zero-shot) ↑ higher is better · zero-shot

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.752 [0.656, 0.849]	100
CASP3_HUMAN_Roychowdhury_2020	0.878 [0.807, 0.943]	100
CCR5_HUMAN_Gill_2023	0.686 [0.578, 0.789]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.822 [0.735, 0.896]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.711 [0.615, 0.813]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.635 [0.499, 0.761]	100
MLAC_ECOLI_MacRae_2023	0.721 [0.621, 0.811]	100
MTH3_HAEAE_RockahShmuel_2015	0.813 [0.686, 0.912]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.901 [0.837, 0.955]	100
SPG1_STRSG_Olson_2014	0.827 [0.713, 0.928]	100

0.775 best

[0.741, 0.803]

1000

1 / 11

0.775

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

MCC (zero-shot) ↑ higher is better · zero-shot

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.420 [0.200, 0.581]	100
CASP3_HUMAN_Roychowdhury_2020	0.581 [0.403, 0.740]	100
CCR5_HUMAN_Gill_2023	0.240 [0.037, 0.424]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.482 [0.260, 0.623]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.218 [0.044, 0.433]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.223 [0.006, 0.406]	100
MLAC_ECOLI_MacRae_2023	0.240 [0.080, 0.461]	100
MTH3_HAEAE_RockahShmuel_2015	0.281 [0.104, 0.485]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.572 [0.378, 0.729]	100
SPG1_STRSG_Olson_2014	0.346 [0.175, 0.469]	100

0.360

[0.299, 0.414]

1000

2 / 11

0.370

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

NDCG, top 10 % (zero-shot) ↑ higher is better · zero-shot

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.818 [0.708, 0.940]	100
CASP3_HUMAN_Roychowdhury_2020	0.858 [0.751, 0.912]	100
CCR5_HUMAN_Gill_2023	0.819 [0.656, 0.874]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.862 [0.761, 0.924]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.823 [0.754, 0.904]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.529 [0.366, 0.690]	100
MLAC_ECOLI_MacRae_2023	0.842 [0.753, 0.907]	100
MTH3_HAEAE_RockahShmuel_2015	0.681 [0.492, 0.828]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.886 [0.801, 0.951]	100
SPG1_STRSG_Olson_2014	0.820 [0.700, 0.939]	100

0.794

[0.750, 0.824]

1000

5 / 11

0.814

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

Top-10 % recall (zero-shot) ↑ higher is better · zero-shot

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.200 [0.000, 0.500]	100
CASP3_HUMAN_Roychowdhury_2020	0.200 [0.000, 0.456]	100
CCR5_HUMAN_Gill_2023	0.000 [0.000, 0.100]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.400 [0.091, 0.700]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.000 [0.000, 0.100]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.100 [0.000, 0.364]	100
MLAC_ECOLI_MacRae_2023	0.000 [0.000, 0.250]	100
MTH3_HAEAE_RockahShmuel_2015	0.091 [0.000, 0.400]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.400 [0.083, 0.700]	100
SPG1_STRSG_Olson_2014	0.300 [0.000, 0.500]	100

0.169

[0.096, 0.243]

1000

5 / 11

0.196

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

Spearman ρ (ridge probe) ↑ higher is better · probe

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.678 [0.552, 0.775]	100
CASP3_HUMAN_Roychowdhury_2020	0.640 [0.516, 0.738]	100
CCR5_HUMAN_Gill_2023	0.211 [0.020, 0.379]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.697 [0.559, 0.786]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.709 [0.555, 0.831]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.536 [0.357, 0.695]	100
MLAC_ECOLI_MacRae_2023	0.482 [0.315, 0.629]	100
MTH3_HAEAE_RockahShmuel_2015	0.571 [0.416, 0.694]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.883 [0.825, 0.921]	100
SPG1_STRSG_Olson_2014	0.784 [0.677, 0.853]	100

0.619

[0.570, 0.657]

1000

4 / 17

0.661

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

AUC (ridge probe) ↑ higher is better · probe

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.791 [0.699, 0.880]	100
CASP3_HUMAN_Roychowdhury_2020	0.840 [0.763, 0.911]	100
CCR5_HUMAN_Gill_2023	0.630 [0.522, 0.732]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.858 [0.777, 0.924]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.948 [0.899, 0.983]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.805 [0.696, 0.905]	100
MLAC_ECOLI_MacRae_2023	0.772 [0.673, 0.863]	100
MTH3_HAEAE_RockahShmuel_2015	0.763 [0.640, 0.867]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.970 [0.931, 0.998]	100
SPG1_STRSG_Olson_2014	0.963 [0.925, 0.992]	100

0.834

[0.808, 0.860]

1000

4 / 17

0.848

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

MCC (ridge probe) ↑ higher is better · probe

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.500 [0.266, 0.660]	100
CASP3_HUMAN_Roychowdhury_2020	0.541 [0.304, 0.682]	100
CCR5_HUMAN_Gill_2023	0.120 [−0.060, 0.342]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.562 [0.360, 0.725]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.611 [0.456, 0.742]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.401 [0.208, 0.559]	100
MLAC_ECOLI_MacRae_2023	0.480 [0.267, 0.631]	100
MTH3_HAEAE_RockahShmuel_2015	0.328 [0.134, 0.507]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.776 [0.601, 0.886]	100
SPG1_STRSG_Olson_2014	0.403 [0.295, 0.521]	100

0.472 best

[0.404, 0.513]

1000

1 / 17

0.472

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

NDCG, top 10 % (ridge probe) ↑ higher is better · probe

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.835 [0.725, 0.962]	100
CASP3_HUMAN_Roychowdhury_2020	0.854 [0.780, 0.918]	100
CCR5_HUMAN_Gill_2023	0.830 [0.617, 0.894]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.828 [0.772, 0.906]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.930 [0.892, 0.967]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.738 [0.578, 0.881]	100
MLAC_ECOLI_MacRae_2023	0.871 [0.799, 0.931]	100
MTH3_HAEAE_RockahShmuel_2015	0.747 [0.604, 0.873]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.908 [0.858, 0.967]	100
SPG1_STRSG_Olson_2014	0.923 [0.869, 0.966]	100

0.846

[0.813, 0.876]

1000

2 / 17

0.860

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

Top-10 % recall (ridge probe) ↑ higher is better · probe

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.500 [0.100, 0.700]	100
CASP3_HUMAN_Roychowdhury_2020	0.200 [0.000, 0.600]	100
CCR5_HUMAN_Gill_2023	0.100 [0.000, 0.333]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.100 [0.000, 0.500]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.600 [0.182, 0.800]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.300 [0.100, 0.636]	100
MLAC_ECOLI_MacRae_2023	0.000 [0.000, 0.300]	100
MTH3_HAEAE_RockahShmuel_2015	0.182 [0.000, 0.455]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.400 [0.100, 0.800]	100
SPG1_STRSG_Olson_2014	0.400 [0.091, 0.700]	100

0.278

[0.209, 0.380]

1000

8 / 17

0.360

reproduced

overlap

Train/test overlap declared by the evaluator: protein-ml/uniref50

Board

Secondary structure (Q3) protein-ml/secondary-structure-q3

Drawn from: tape

atlas-n100-v2

Metric	Value	95% CI	n	Rank	Best in catalogue	Source	Overlap	Board
Q3 accuracy (per-residue probe) ↑ higher is better · probe	0.797	[0.783, 0.810]	100	4 / 17	0.813	reproduced	overlap Train/test overlap declared by the evaluator: protein-ml/uniref50	Board
Q8 accuracy (per-residue probe) ↑ higher is better · probe	0.654	[0.635, 0.672]	100	3 / 17	0.663	reproduced	overlap Train/test overlap declared by the evaluator: protein-ml/uniref50	Board