Models / protein-ml / progen2-small / benchmarks

protein-ml/progen2-small

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	0.226 [−0.202, 0.600]	25
MET_HUMAN_Estevam_2023	0.237 [−0.109, 0.546]	25
S22A1_HUMAN_Yee_2023_activity	0.458 [0.033, 0.765]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.433 [0.019, 0.706]	25

0.338

[0.119, 0.502]

100

6 / 13

0.476

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

Board

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

By assay: 4 levels

assay	Value [95% CI]	n
CCR5_HUMAN_Gill_2023	0.532 [0.287, 0.773]	25
MET_HUMAN_Estevam_2023	0.701 [0.480, 0.890]	25
S22A1_HUMAN_Yee_2023_activity	0.720 [0.500, 0.896]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.694 [0.449, 0.893]	25

0.662

[0.542, 0.771]

100

8 / 13

0.759

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

Board

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

By assay: 4 levels

assay	Value [95% CI]	n
CCR5_HUMAN_Gill_2023	0.038 [−0.439, 0.447]	25
MET_HUMAN_Estevam_2023	0.116 [−0.263, 0.603]	25
S22A1_HUMAN_Yee_2023_activity	0.294 [−0.133, 0.721]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.113 [−0.272, 0.634]	25

0.141

[−0.048, 0.389]

100

12 / 13

0.426

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

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, 0.666]	25
MET_HUMAN_Estevam_2023	0.877 [0.749, 0.969]	25
S22A1_HUMAN_Yee_2023_activity	0.958 [0.811, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.209 [0.079, 0.605]	25

0.586

[0.490, 0.731]

100

—

0.635

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

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, 0.000]	25
MET_HUMAN_Estevam_2023	0.000 [0.000, 0.667]	25
S22A1_HUMAN_Yee_2023_activity	0.667 [0.000, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.000 [0.000, 0.333]	25

0.167 best

[0.000, 0.375]

100

—

0.167

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

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	0.226 [−0.202, 0.600]	25
MET_HUMAN_Estevam_2023	0.237 [−0.109, 0.546]	25
S22A1_HUMAN_Yee_2023_activity	0.458 [0.033, 0.765]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.433 [0.019, 0.706]	25

0.338

[0.119, 0.502]

100

6 / 13

0.476

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

Board

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

By assay: 4 levels

assay	Value [95% CI]	n
CCR5_HUMAN_Gill_2023	0.532 [0.287, 0.773]	25
MET_HUMAN_Estevam_2023	0.701 [0.480, 0.890]	25
S22A1_HUMAN_Yee_2023_activity	0.720 [0.500, 0.896]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.694 [0.449, 0.893]	25

0.662

[0.542, 0.771]

100

8 / 13

0.759

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

Board

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

By assay: 4 levels

assay	Value [95% CI]	n
CCR5_HUMAN_Gill_2023	0.038 [−0.439, 0.447]	25
MET_HUMAN_Estevam_2023	0.116 [−0.263, 0.603]	25
S22A1_HUMAN_Yee_2023_activity	0.294 [−0.133, 0.721]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.113 [−0.272, 0.634]	25

0.141

[−0.048, 0.389]

100

12 / 13

0.426

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

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, 0.666]	25
MET_HUMAN_Estevam_2023	0.877 [0.749, 0.969]	25
S22A1_HUMAN_Yee_2023_activity	0.958 [0.811, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.209 [0.079, 0.605]	25

0.586

[0.490, 0.731]

100

—

0.635

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

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, 0.000]	25
MET_HUMAN_Estevam_2023	0.000 [0.000, 0.667]	25
S22A1_HUMAN_Yee_2023_activity	0.667 [0.000, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.000 [0.000, 0.333]	25

0.167 best

[0.000, 0.375]

100

—

0.167

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

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	0.188 [−0.211, 0.511]	25
CCR5_HUMAN_Gill_2023	0.245 [−0.186, 0.609]	25
OPSD_HUMAN_Wan_2019	0.476 [0.071, 0.725]	25
S22A1_HUMAN_Yee_2023_activity	0.456 [0.031, 0.761]	25

0.341

[0.136, 0.509]

100

5 / 13

0.463

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

Board

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

By assay: 4 levels

assay	Value [95% CI]	n
ADRB2_HUMAN_Jones_2020	0.539 [0.285, 0.787]	25
CCR5_HUMAN_Gill_2023	0.545 [0.294, 0.776]	25
OPSD_HUMAN_Wan_2019	0.868 [0.683, 1.000]	25
S22A1_HUMAN_Yee_2023_activity	0.720 [0.500, 0.896]	25

0.668

[0.555, 0.766]

100

6 / 13

0.751

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

Board

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

By assay: 4 levels

assay	Value [95% CI]	n
ADRB2_HUMAN_Jones_2020	−0.206 [−0.510, 0.439]	25
CCR5_HUMAN_Gill_2023	0.038 [−0.439, 0.447]	25
OPSD_HUMAN_Wan_2019	0.280 [−0.014, 0.840]	25
S22A1_HUMAN_Yee_2023_activity	0.294 [−0.133, 0.721]	25

0.102

[−0.045, 0.380]

100

10 / 13

0.426

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

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	0.746 [0.365, 1.000]	25
CCR5_HUMAN_Gill_2023	0.548 [0.000, 0.775]	25
OPSD_HUMAN_Wan_2019	0.593 [0.535, 0.877]	25
S22A1_HUMAN_Yee_2023_activity	0.949 [0.811, 1.000]	25

0.709 best

[0.563, 0.819]

100

—

0.709

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

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	0.333 [0.000, 1.000]	25
CCR5_HUMAN_Gill_2023	0.000 [0.000, 0.144]	25
OPSD_HUMAN_Wan_2019	0.000 [0.000, 1.000]	25
S22A1_HUMAN_Yee_2023_activity	0.667 [0.000, 1.000]	25

0.250 best

[0.000, 0.542]

100

—

0.250

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

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	0.237 [−0.109, 0.546]	25
MK01_HUMAN_Brenan_2016	0.313 [−0.140, 0.677]	25
SCN5A_HUMAN_Glazer_2019	0.150 [−0.352, 0.547]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.433 [0.019, 0.706]	25

0.283

[0.079, 0.460]

100

2 / 13

0.362

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

Board

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

By assay: 4 levels

assay	Value [95% CI]	n
MET_HUMAN_Estevam_2023	0.701 [0.480, 0.890]	25
MK01_HUMAN_Brenan_2016	0.765 [0.479, 0.984]	25
SCN5A_HUMAN_Glazer_2019	0.527 [0.272, 0.756]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.694 [0.449, 0.893]	25

0.672

[0.542, 0.784]

100

7 / 13

0.722

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

Board

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

By assay: 4 levels

assay	Value [95% CI]	n
MET_HUMAN_Estevam_2023	0.116 [−0.263, 0.603]	25
MK01_HUMAN_Brenan_2016	0.316 [−0.115, 0.690]	25
SCN5A_HUMAN_Glazer_2019	0.033 [−0.395, 0.408]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.113 [−0.272, 0.634]	25

0.145

[−0.061, 0.392]

100

11 / 13

0.310

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

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	0.877 [0.749, 0.969]	25
MK01_HUMAN_Brenan_2016	0.409 [0.000, 0.919]	25
SCN5A_HUMAN_Glazer_2019	0.800 [0.289, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.209 [0.079, 0.605]	25

0.574

[0.402, 0.793]

100

—

0.725

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

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	0.000 [0.000, 0.667]	25
MK01_HUMAN_Brenan_2016	0.000 [0.000, 0.600]	25
SCN5A_HUMAN_Glazer_2019	0.667 [0.000, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.000 [0.000, 0.333]	25

0.167

[0.000, 0.396]

100

—

0.250

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

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	0.711	[0.651, 0.794]	100	3 / 17	0.844	reproduced	unknown	Board
Fmax (biological process, probe) ↑ higher is better · probe	0.445	[0.356, 0.558]	100	11 / 17	0.538	reproduced	unknown	Board
Fmax (cellular component, probe) ↑ higher is better · probe	0.713	[0.660, 0.773]	100	16 / 17	0.756	reproduced	unknown	Board
Micro AUPR (molecular function, probe) ↑ higher is better · probe · secondary	0.713	[0.631, 0.794]	100	—	0.875	reproduced	unknown	Board
Smin (molecular function, probe) ↓ lower is better · bits · probe · secondary	2.964	[2.099, 3.740]	100	—	1.727	reproduced	unknown	Board
Micro AUPR (biological process, probe) ↑ higher is better · probe · secondary	0.409	[0.271, 0.566]	100	—	0.595	reproduced	unknown	Board
Smin (biological process, probe) ↓ lower is better · bits · probe · secondary	6.521	[4.596, 7.995]	100	—	4.842	reproduced	unknown	Board
Micro AUPR (cellular component, probe) ↑ higher is better · probe · secondary	0.677	[0.607, 0.745]	100	—	0.758	reproduced	unknown	Board
Smin (cellular component, probe) ↓ lower is better · bits · probe · secondary	3.458	[2.724, 4.111]	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	0.811	[0.794, 0.829]	1728	17 / 17	0.888	reproduced	overlap Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90	Board
MCC (logistic probe) ↑ higher is better · probe · secondary	0.620	[0.585, 0.657]	1728	—	0.771	reproduced	overlap Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90	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	0.116	[−0.113, 0.314]	100	12 / 12	0.410	reproduced	unknown	Board
Spearman ρ (ridge probe) ↑ higher is better · probe	0.034	[−0.142, 0.197]	100	16 / 17	0.297	reproduced	unknown	Board
RMSE, kcal/mol (ridge probe) ↓ lower is better · kcal/mol · probe	1.114	[0.884, 1.320]	100	10 / 17	1.095	reproduced	unknown	Board
Pearson r (ridge probe) ↑ higher is better · probe · secondary	0.045	[−0.128, 0.193]	100	—	0.349	reproduced	unknown	Board
MAE, kcal/mol (ridge probe) ↓ lower is better · kcal/mol · probe · secondary	0.851	[0.706, 1.006]	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	0.585	[0.571, 0.599]	2742	17 / 17	0.801	reproduced	overlap Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90	Board
Macro F1 (logistic probe) ↑ higher is better · probe	0.387	[0.368, 0.404]	2742	17 / 17	0.640	reproduced	overlap Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90	Board
MCC (logistic probe) ↑ higher is better · probe · secondary	0.491	[0.472, 0.509]	2742	—	0.758	reproduced	overlap Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90	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	0.524 [0.356, 0.670]	100
CASP3_HUMAN_Roychowdhury_2020	0.510 [0.356, 0.636]	100
CCR5_HUMAN_Gill_2023	0.326 [0.149, 0.505]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.586 [0.436, 0.699]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.315 [0.132, 0.483]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.142 [−0.064, 0.348]	100
MLAC_ECOLI_MacRae_2023	0.310 [0.145, 0.459]	100
MTH3_HAEAE_RockahShmuel_2015	0.516 [0.345, 0.647]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.728 [0.591, 0.828]	100
SPG1_STRSG_Olson_2014	0.038 [−0.186, 0.240]	100

0.399

[0.340, 0.448]

1000

8 / 11

0.434

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

Board

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

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.756 [0.655, 0.853]	100
CASP3_HUMAN_Roychowdhury_2020	0.781 [0.690, 0.877]	100
CCR5_HUMAN_Gill_2023	0.676 [0.573, 0.780]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.793 [0.692, 0.870]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.652 [0.529, 0.767]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.658 [0.532, 0.775]	100
MLAC_ECOLI_MacRae_2023	0.739 [0.645, 0.828]	100
MTH3_HAEAE_RockahShmuel_2015	0.791 [0.659, 0.899]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.932 [0.875, 0.975]	100
SPG1_STRSG_Olson_2014	0.698 [0.536, 0.829]	100

0.748

[0.711, 0.781]

1000

7 / 11

0.775

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

Board

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

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.420 [0.260, 0.621]	100
CASP3_HUMAN_Roychowdhury_2020	0.461 [0.219, 0.642]	100
CCR5_HUMAN_Gill_2023	0.240 [0.020, 0.422]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.361 [0.179, 0.580]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.218 [0.038, 0.439]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.223 [0.023, 0.394]	100
MLAC_ECOLI_MacRae_2023	0.320 [0.155, 0.521]	100
MTH3_HAEAE_RockahShmuel_2015	0.375 [0.178, 0.535]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.694 [0.551, 0.840]	100
SPG1_STRSG_Olson_2014	0.231 [0.022, 0.382]	100

0.354

[0.296, 0.416]

1000

4 / 11

0.370

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

Board

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

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.744 [0.632, 0.884]	100
CASP3_HUMAN_Roychowdhury_2020	0.742 [0.669, 0.847]	100
CCR5_HUMAN_Gill_2023	0.828 [0.651, 0.871]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.872 [0.799, 0.962]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.838 [0.744, 0.916]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.533 [0.342, 0.679]	100
MLAC_ECOLI_MacRae_2023	0.826 [0.663, 0.899]	100
MTH3_HAEAE_RockahShmuel_2015	0.565 [0.326, 0.837]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.877 [0.814, 0.950]	100
SPG1_STRSG_Olson_2014	0.875 [0.765, 0.944]	100

0.770

[0.723, 0.812]

1000

7 / 11

0.814

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

Board

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

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.000 [0.000, 0.300]	100
CASP3_HUMAN_Roychowdhury_2020	0.100 [0.000, 0.300]	100
CCR5_HUMAN_Gill_2023	0.100 [0.000, 0.300]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.400 [0.152, 0.700]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.100 [0.000, 0.364]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.000 [0.000, 0.357]	100
MLAC_ECOLI_MacRae_2023	0.200 [0.000, 0.455]	100
MTH3_HAEAE_RockahShmuel_2015	0.182 [0.000, 0.500]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.400 [0.100, 0.700]	100
SPG1_STRSG_Olson_2014	0.300 [0.000, 0.500]	100

0.178

[0.107, 0.261]

1000

3 / 11

0.196

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

Board

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

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.690 [0.547, 0.795]	100
CASP3_HUMAN_Roychowdhury_2020	0.451 [0.283, 0.590]	100
CCR5_HUMAN_Gill_2023	0.431 [0.263, 0.578]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.531 [0.370, 0.666]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.689 [0.539, 0.800]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.607 [0.458, 0.728]	100
MLAC_ECOLI_MacRae_2023	0.362 [0.159, 0.531]	100
MTH3_HAEAE_RockahShmuel_2015	0.393 [0.212, 0.541]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.877 [0.807, 0.921]	100
SPG1_STRSG_Olson_2014	0.771 [0.659, 0.853]	100

0.580

[0.528, 0.617]

1000

9 / 17

0.661

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

Board

AUC (ridge probe) ↑ higher is better · probe

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.821 [0.725, 0.909]	100
CASP3_HUMAN_Roychowdhury_2020	0.742 [0.636, 0.842]	100
CCR5_HUMAN_Gill_2023	0.702 [0.604, 0.800]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.756 [0.655, 0.846]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.893 [0.799, 0.961]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.846 [0.755, 0.924]	100
MLAC_ECOLI_MacRae_2023	0.710 [0.604, 0.809]	100
MTH3_HAEAE_RockahShmuel_2015	0.709 [0.586, 0.823]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.961 [0.921, 0.991]	100
SPG1_STRSG_Olson_2014	0.914 [0.814, 0.982]	100

0.806

[0.774, 0.834]

1000

9 / 17

0.848

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

Board

MCC (ridge probe) ↑ higher is better · probe

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.540 [0.360, 0.740]	100
CASP3_HUMAN_Roychowdhury_2020	0.421 [0.182, 0.565]	100
CCR5_HUMAN_Gill_2023	0.280 [0.121, 0.500]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.401 [0.200, 0.602]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.567 [0.382, 0.684]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.445 [0.269, 0.603]	100
MLAC_ECOLI_MacRae_2023	0.280 [0.120, 0.521]	100
MTH3_HAEAE_RockahShmuel_2015	0.328 [0.123, 0.485]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.776 [0.595, 0.869]	100
SPG1_STRSG_Olson_2014	0.346 [0.185, 0.469]	100

0.438

[0.381, 0.491]

1000

7 / 17

0.472

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

Board

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

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.771 [0.673, 0.920]	100
CASP3_HUMAN_Roychowdhury_2020	0.745 [0.639, 0.861]	100
CCR5_HUMAN_Gill_2023	0.811 [0.629, 0.894]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.764 [0.615, 0.908]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.894 [0.861, 0.958]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.738 [0.590, 0.878]	100
MLAC_ECOLI_MacRae_2023	0.805 [0.663, 0.917]	100
MTH3_HAEAE_RockahShmuel_2015	0.518 [0.311, 0.812]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.954 [0.914, 0.984]	100
SPG1_STRSG_Olson_2014	0.938 [0.895, 0.980]	100

0.794

[0.755, 0.835]

1000

13 / 17

0.860

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

Board

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

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.200 [0.000, 0.545]	100
CASP3_HUMAN_Roychowdhury_2020	0.300 [0.000, 0.583]	100
CCR5_HUMAN_Gill_2023	0.200 [0.000, 0.500]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.400 [0.000, 0.700]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.400 [0.100, 0.700]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.500 [0.100, 0.700]	100
MLAC_ECOLI_MacRae_2023	0.100 [0.000, 0.400]	100
MTH3_HAEAE_RockahShmuel_2015	0.091 [0.000, 0.333]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.500 [0.231, 0.818]	100
SPG1_STRSG_Olson_2014	0.400 [0.091, 0.727]	100

0.309

[0.207, 0.382]

1000

4 / 17

0.360

reproduced

overlap

Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90

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	0.625	[0.609, 0.639]	100	16 / 17	0.813	reproduced	overlap Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90	Board
Q8 accuracy (per-residue probe) ↑ higher is better · probe	0.474	[0.456, 0.491]	100	16 / 17	0.663	reproduced	overlap Train/test overlap declared by the evaluator: BFD, protein-ml/uniref90	Board