Benchmarks · protein-ml/esm2-t6-8m-ur50d

ProteinGym protein-ml/proteingym

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

n

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	0.289 [−0.149, 0.653]	25
MET_HUMAN_Estevam_2023	0.438 [0.106, 0.703]	25
S22A1_HUMAN_Yee_2023_activity	0.408 [−0.004, 0.723]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.251 [−0.116, 0.603]	25

0.347

[0.130, 0.506]

100

5 / 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	0.667 [0.433, 0.872]	25
MET_HUMAN_Estevam_2023	0.753 [0.545, 0.924]	25
S22A1_HUMAN_Yee_2023_activity	0.720 [0.487, 0.919]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.681 [0.455, 0.877]	25

0.705

[0.588, 0.808]

100

5 / 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 [−0.206, 0.624]	25
MET_HUMAN_Estevam_2023	0.439 [−0.122, 0.714]	25
S22A1_HUMAN_Yee_2023_activity	0.458 [−0.116, 0.690]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.280 [−0.220, 0.614]	25

0.344

[0.057, 0.473]

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
CCR5_HUMAN_Gill_2023	0.548 [0.000, 0.770]	25
MET_HUMAN_Estevam_2023	0.789 [0.630, 0.985]	25
S22A1_HUMAN_Yee_2023_activity	0.653 [0.192, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.097 [0.059, 1.000]	25

0.522

[0.373, 0.796]

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, 0.502]	25
MET_HUMAN_Estevam_2023	0.000 [0.000, 0.751]	25
S22A1_HUMAN_Yee_2023_activity	0.333 [0.000, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.333 [0.000, 1.000]	25

0.167 best

[0.000, 0.500]

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

n

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	0.289 [−0.149, 0.653]	25
MET_HUMAN_Estevam_2023	0.438 [0.106, 0.703]	25
S22A1_HUMAN_Yee_2023_activity	0.408 [−0.004, 0.723]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.251 [−0.116, 0.603]	25

0.347

[0.130, 0.506]

100

5 / 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	0.667 [0.433, 0.872]	25
MET_HUMAN_Estevam_2023	0.753 [0.545, 0.924]	25
S22A1_HUMAN_Yee_2023_activity	0.720 [0.487, 0.919]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.681 [0.455, 0.877]	25

0.705

[0.588, 0.808]

100

5 / 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 [−0.206, 0.624]	25
MET_HUMAN_Estevam_2023	0.439 [−0.122, 0.714]	25
S22A1_HUMAN_Yee_2023_activity	0.458 [−0.116, 0.690]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.280 [−0.220, 0.614]	25

0.344

[0.057, 0.473]

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
CCR5_HUMAN_Gill_2023	0.548 [0.000, 0.770]	25
MET_HUMAN_Estevam_2023	0.789 [0.630, 0.985]	25
S22A1_HUMAN_Yee_2023_activity	0.653 [0.192, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.097 [0.059, 1.000]	25

0.522

[0.373, 0.796]

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, 0.502]	25
MET_HUMAN_Estevam_2023	0.000 [0.000, 0.751]	25
S22A1_HUMAN_Yee_2023_activity	0.333 [0.000, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.333 [0.000, 1.000]	25

0.167 best

[0.000, 0.500]

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

n

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	0.041 [−0.352, 0.437]	25
CCR5_HUMAN_Gill_2023	0.289 [−0.149, 0.653]	25
OPSD_HUMAN_Wan_2019	−0.080 [−0.464, 0.357]	25
S22A1_HUMAN_Yee_2023_activity	0.408 [−0.004, 0.723]	25

0.164

[−0.038, 0.362]

100

11 / 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	0.506 [0.271, 0.753]	25
CCR5_HUMAN_Gill_2023	0.667 [0.433, 0.872]	25
OPSD_HUMAN_Wan_2019	0.479 [0.243, 0.720]	25
S22A1_HUMAN_Yee_2023_activity	0.720 [0.487, 0.919]	25

0.593

[0.483, 0.711]

100

11 / 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	−0.045 [−0.442, 0.428]	25
CCR5_HUMAN_Gill_2023	0.199 [−0.206, 0.624]	25
OPSD_HUMAN_Wan_2019	−0.220 [−0.603, 0.292]	25
S22A1_HUMAN_Yee_2023_activity	0.458 [−0.116, 0.690]	25

0.098

[−0.127, 0.329]

100

12 / 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	0.254 [0.000, 0.747]	25
CCR5_HUMAN_Gill_2023	0.548 [0.000, 0.770]	25
OPSD_HUMAN_Wan_2019	0.015 [0.002, 0.444]	25
S22A1_HUMAN_Yee_2023_activity	0.653 [0.192, 1.000]	25

0.368

[0.223, 0.588]

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, 0.667]	25
CCR5_HUMAN_Gill_2023	0.000 [0.000, 0.502]	25
OPSD_HUMAN_Wan_2019	0.000 [0.000, 0.667]	25
S22A1_HUMAN_Yee_2023_activity	0.333 [0.000, 1.000]	25

0.083

[0.000, 0.375]

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

n

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	0.438 [0.106, 0.703]	25
MK01_HUMAN_Brenan_2016	0.341 [−0.106, 0.694]	25
SCN5A_HUMAN_Glazer_2019	0.092 [−0.356, 0.496]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.251 [−0.116, 0.603]	25

0.281

[0.072, 0.455]

100

3 / 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	0.753 [0.545, 0.924]	25
MK01_HUMAN_Brenan_2016	0.721 [0.452, 0.942]	25
SCN5A_HUMAN_Glazer_2019	0.513 [0.260, 0.740]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.681 [0.455, 0.877]	25

0.667

[0.540, 0.782]

100

9 / 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	0.439 [−0.122, 0.714]	25
MK01_HUMAN_Brenan_2016	0.316 [−0.116, 0.659]	25
SCN5A_HUMAN_Glazer_2019	0.033 [−0.428, 0.447]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.280 [−0.220, 0.614]	25

0.267

[−0.020, 0.428]

100

5 / 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	0.789 [0.630, 0.985]	25
MK01_HUMAN_Brenan_2016	0.399 [0.000, 0.923]	25
SCN5A_HUMAN_Glazer_2019	0.233 [0.069, 0.977]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.097 [0.059, 1.000]	25

0.380

[0.256, 0.749]

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	0.000 [0.000, 0.751]	25
MK01_HUMAN_Brenan_2016	0.000 [0.000, 1.000]	25
SCN5A_HUMAN_Glazer_2019	0.333 [0.000, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.333 [0.000, 1.000]	25

0.167

[0.000, 0.563]

100

—

0.250

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

n

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	0.494 [0.316, 0.645]	100
CASP3_HUMAN_Roychowdhury_2020	0.341 [0.155, 0.503]	100
CCR5_HUMAN_Gill_2023	0.262 [0.078, 0.454]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.537 [0.356, 0.670]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.352 [0.148, 0.524]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.043 [−0.160, 0.240]	100
MLAC_ECOLI_MacRae_2023	−0.009 [−0.207, 0.188]	100
MTH3_HAEAE_RockahShmuel_2015	0.206 [0.010, 0.388]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.734 [0.605, 0.825]	100
SPG1_STRSG_Olson_2014	0.196 [−0.020, 0.397]	100

0.316

[0.254, 0.369]

1000

10 / 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	0.712 [0.610, 0.810]	100
CASP3_HUMAN_Roychowdhury_2020	0.703 [0.592, 0.806]	100
CCR5_HUMAN_Gill_2023	0.635 [0.528, 0.746]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.747 [0.640, 0.833]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.750 [0.632, 0.852]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.593 [0.470, 0.709]	100
MLAC_ECOLI_MacRae_2023	0.544 [0.428, 0.659]	100
MTH3_HAEAE_RockahShmuel_2015	0.651 [0.508, 0.781]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.932 [0.880, 0.976]	100
SPG1_STRSG_Olson_2014	0.762 [0.618, 0.886]	100

0.703

[0.667, 0.739]

1000

10 / 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	0.260 [0.066, 0.480]	100
CASP3_HUMAN_Roychowdhury_2020	0.341 [0.098, 0.510]	100
CCR5_HUMAN_Gill_2023	0.200 [0.020, 0.422]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.281 [0.100, 0.479]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.349 [0.145, 0.504]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.089 [−0.119, 0.281]	100
MLAC_ECOLI_MacRae_2023	0.080 [−0.160, 0.280]	100
MTH3_HAEAE_RockahShmuel_2015	0.094 [−0.074, 0.312]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.653 [0.540, 0.821]	100
SPG1_STRSG_Olson_2014	0.231 [0.043, 0.404]	100

0.258

[0.206, 0.327]

1000

10 / 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	0.785 [0.653, 0.917]	100
CASP3_HUMAN_Roychowdhury_2020	0.697 [0.639, 0.781]	100
CCR5_HUMAN_Gill_2023	0.806 [0.594, 0.875]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.813 [0.690, 0.898]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.828 [0.746, 0.907]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.582 [0.418, 0.709]	100
MLAC_ECOLI_MacRae_2023	0.772 [0.552, 0.904]	100
MTH3_HAEAE_RockahShmuel_2015	0.406 [0.212, 0.582]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.891 [0.806, 0.964]	100
SPG1_STRSG_Olson_2014	0.873 [0.724, 0.950]	100

0.745

[0.695, 0.779]

1000

10 / 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	0.100 [0.000, 0.500]	100
CASP3_HUMAN_Roychowdhury_2020	0.000 [0.000, 0.201]	100
CCR5_HUMAN_Gill_2023	0.100 [0.000, 0.400]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.300 [0.091, 0.600]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.100 [0.000, 0.364]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.200 [0.000, 0.500]	100
MLAC_ECOLI_MacRae_2023	0.200 [0.000, 0.500]	100
MTH3_HAEAE_RockahShmuel_2015	0.000 [0.000, 0.101]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.500 [0.100, 0.727]	100
SPG1_STRSG_Olson_2014	0.200 [0.000, 0.500]	100

0.170

[0.098, 0.257]

1000

4 / 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	0.578 [0.405, 0.709]	100
CASP3_HUMAN_Roychowdhury_2020	0.122 [−0.089, 0.333]	100
CCR5_HUMAN_Gill_2023	0.250 [0.064, 0.419]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.390 [0.195, 0.556]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.668 [0.522, 0.789]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.455 [0.264, 0.607]	100
MLAC_ECOLI_MacRae_2023	0.316 [0.113, 0.485]	100
MTH3_HAEAE_RockahShmuel_2015	0.309 [0.116, 0.471]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.890 [0.839, 0.921]	100
SPG1_STRSG_Olson_2014	0.784 [0.666, 0.863]	100

0.476

[0.418, 0.523]

1000

16 / 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	0.800 [0.713, 0.884]	100
CASP3_HUMAN_Roychowdhury_2020	0.531 [0.413, 0.647]	100
CCR5_HUMAN_Gill_2023	0.621 [0.513, 0.719]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.675 [0.556, 0.772]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.874 [0.783, 0.950]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.762 [0.641, 0.875]	100
MLAC_ECOLI_MacRae_2023	0.660 [0.552, 0.762]	100
MTH3_HAEAE_RockahShmuel_2015	0.672 [0.551, 0.789]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.956 [0.918, 0.989]	100
SPG1_STRSG_Olson_2014	0.941 [0.889, 0.978]	100

0.749

[0.718, 0.781]

1000

16 / 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	0.420 [0.221, 0.600]	100
CASP3_HUMAN_Roychowdhury_2020	0.060 [−0.160, 0.260]	100
CCR5_HUMAN_Gill_2023	0.080 [−0.081, 0.338]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.281 [0.043, 0.484]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.393 [0.193, 0.579]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.356 [0.178, 0.531]	100
MLAC_ECOLI_MacRae_2023	0.280 [0.041, 0.461]	100
MTH3_HAEAE_RockahShmuel_2015	0.234 [−0.013, 0.381]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.694 [0.555, 0.841]	100
SPG1_STRSG_Olson_2014	0.403 [0.295, 0.521]	100

0.320

[0.257, 0.378]

1000

16 / 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	0.784 [0.680, 0.921]	100
CASP3_HUMAN_Roychowdhury_2020	0.629 [0.389, 0.844]	100
CCR5_HUMAN_Gill_2023	0.808 [0.608, 0.879]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.588 [0.461, 0.718]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.924 [0.877, 0.965]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.652 [0.453, 0.812]	100
MLAC_ECOLI_MacRae_2023	0.856 [0.760, 0.934]	100
MTH3_HAEAE_RockahShmuel_2015	0.657 [0.439, 0.814]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.935 [0.906, 0.983]	100
SPG1_STRSG_Olson_2014	0.920 [0.850, 0.971]	100

0.775

[0.727, 0.816]

1000

15 / 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	0.400 [0.000, 0.546]	100
CASP3_HUMAN_Roychowdhury_2020	0.200 [0.000, 0.500]	100
CCR5_HUMAN_Gill_2023	0.100 [0.000, 0.455]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.000 [0.000, 0.100]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.200 [0.000, 0.600]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.100 [0.000, 0.364]	100
MLAC_ECOLI_MacRae_2023	0.100 [0.000, 0.500]	100
MTH3_HAEAE_RockahShmuel_2015	0.091 [0.000, 0.400]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.500 [0.200, 0.900]	100
SPG1_STRSG_Olson_2014	0.500 [0.182, 0.701]	100

0.219

[0.153, 0.318]

1000

12 / 17

0.360

reproduced

overlap

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

Board

ProteinGym_v1 protein-ml/proteingym-v1

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

n

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	0.289 [−0.149, 0.653]	25
MET_HUMAN_Estevam_2023	0.438 [0.106, 0.703]	25
S22A1_HUMAN_Yee_2023_activity	0.408 [−0.004, 0.723]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.251 [−0.116, 0.603]	25

0.347

[0.130, 0.506]

100

5 / 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	0.667 [0.433, 0.872]	25
MET_HUMAN_Estevam_2023	0.753 [0.545, 0.924]	25
S22A1_HUMAN_Yee_2023_activity	0.720 [0.487, 0.919]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.681 [0.455, 0.877]	25

0.705

[0.588, 0.808]

100

5 / 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 [−0.206, 0.624]	25
MET_HUMAN_Estevam_2023	0.439 [−0.122, 0.714]	25
S22A1_HUMAN_Yee_2023_activity	0.458 [−0.116, 0.690]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.280 [−0.220, 0.614]	25

0.344

[0.057, 0.473]

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
CCR5_HUMAN_Gill_2023	0.548 [0.000, 0.770]	25
MET_HUMAN_Estevam_2023	0.789 [0.630, 0.985]	25
S22A1_HUMAN_Yee_2023_activity	0.653 [0.192, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.097 [0.059, 1.000]	25

0.522

[0.373, 0.796]

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, 0.502]	25
MET_HUMAN_Estevam_2023	0.000 [0.000, 0.751]	25
S22A1_HUMAN_Yee_2023_activity	0.333 [0.000, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.333 [0.000, 1.000]	25

0.167 best

[0.000, 0.500]

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

n

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	0.289 [−0.149, 0.653]	25
MET_HUMAN_Estevam_2023	0.438 [0.106, 0.703]	25
S22A1_HUMAN_Yee_2023_activity	0.408 [−0.004, 0.723]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.251 [−0.116, 0.603]	25

0.347

[0.130, 0.506]

100

5 / 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	0.667 [0.433, 0.872]	25
MET_HUMAN_Estevam_2023	0.753 [0.545, 0.924]	25
S22A1_HUMAN_Yee_2023_activity	0.720 [0.487, 0.919]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.681 [0.455, 0.877]	25

0.705

[0.588, 0.808]

100

5 / 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 [−0.206, 0.624]	25
MET_HUMAN_Estevam_2023	0.439 [−0.122, 0.714]	25
S22A1_HUMAN_Yee_2023_activity	0.458 [−0.116, 0.690]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.280 [−0.220, 0.614]	25

0.344

[0.057, 0.473]

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
CCR5_HUMAN_Gill_2023	0.548 [0.000, 0.770]	25
MET_HUMAN_Estevam_2023	0.789 [0.630, 0.985]	25
S22A1_HUMAN_Yee_2023_activity	0.653 [0.192, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.097 [0.059, 1.000]	25

0.522

[0.373, 0.796]

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, 0.502]	25
MET_HUMAN_Estevam_2023	0.000 [0.000, 0.751]	25
S22A1_HUMAN_Yee_2023_activity	0.333 [0.000, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.333 [0.000, 1.000]	25

0.167 best

[0.000, 0.500]

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

n

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	0.041 [−0.352, 0.437]	25
CCR5_HUMAN_Gill_2023	0.289 [−0.149, 0.653]	25
OPSD_HUMAN_Wan_2019	−0.080 [−0.464, 0.357]	25
S22A1_HUMAN_Yee_2023_activity	0.408 [−0.004, 0.723]	25

0.164

[−0.038, 0.362]

100

11 / 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	0.506 [0.271, 0.753]	25
CCR5_HUMAN_Gill_2023	0.667 [0.433, 0.872]	25
OPSD_HUMAN_Wan_2019	0.479 [0.243, 0.720]	25
S22A1_HUMAN_Yee_2023_activity	0.720 [0.487, 0.919]	25

0.593

[0.483, 0.711]

100

11 / 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	−0.045 [−0.442, 0.428]	25
CCR5_HUMAN_Gill_2023	0.199 [−0.206, 0.624]	25
OPSD_HUMAN_Wan_2019	−0.220 [−0.603, 0.292]	25
S22A1_HUMAN_Yee_2023_activity	0.458 [−0.116, 0.690]	25

0.098

[−0.127, 0.329]

100

12 / 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	0.254 [0.000, 0.747]	25
CCR5_HUMAN_Gill_2023	0.548 [0.000, 0.770]	25
OPSD_HUMAN_Wan_2019	0.015 [0.002, 0.444]	25
S22A1_HUMAN_Yee_2023_activity	0.653 [0.192, 1.000]	25

0.368

[0.223, 0.588]

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, 0.667]	25
CCR5_HUMAN_Gill_2023	0.000 [0.000, 0.502]	25
OPSD_HUMAN_Wan_2019	0.000 [0.000, 0.667]	25
S22A1_HUMAN_Yee_2023_activity	0.333 [0.000, 1.000]	25

0.083

[0.000, 0.375]

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

n

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	0.438 [0.106, 0.703]	25
MK01_HUMAN_Brenan_2016	0.341 [−0.106, 0.694]	25
SCN5A_HUMAN_Glazer_2019	0.092 [−0.356, 0.496]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.251 [−0.116, 0.603]	25

0.281

[0.072, 0.455]

100

3 / 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	0.753 [0.545, 0.924]	25
MK01_HUMAN_Brenan_2016	0.721 [0.452, 0.942]	25
SCN5A_HUMAN_Glazer_2019	0.513 [0.260, 0.740]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.681 [0.455, 0.877]	25

0.667

[0.540, 0.782]

100

9 / 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	0.439 [−0.122, 0.714]	25
MK01_HUMAN_Brenan_2016	0.316 [−0.116, 0.659]	25
SCN5A_HUMAN_Glazer_2019	0.033 [−0.428, 0.447]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.280 [−0.220, 0.614]	25

0.267

[−0.020, 0.428]

100

5 / 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	0.789 [0.630, 0.985]	25
MK01_HUMAN_Brenan_2016	0.399 [0.000, 0.923]	25
SCN5A_HUMAN_Glazer_2019	0.233 [0.069, 0.977]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.097 [0.059, 1.000]	25

0.380

[0.256, 0.749]

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	0.000 [0.000, 0.751]	25
MK01_HUMAN_Brenan_2016	0.000 [0.000, 1.000]	25
SCN5A_HUMAN_Glazer_2019	0.333 [0.000, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.333 [0.000, 1.000]	25

0.167

[0.000, 0.563]

100

—

0.250

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

n

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	0.494 [0.316, 0.645]	100
CASP3_HUMAN_Roychowdhury_2020	0.341 [0.155, 0.503]	100
CCR5_HUMAN_Gill_2023	0.262 [0.078, 0.454]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.537 [0.356, 0.670]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.352 [0.148, 0.524]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.043 [−0.160, 0.240]	100
MLAC_ECOLI_MacRae_2023	−0.009 [−0.207, 0.188]	100
MTH3_HAEAE_RockahShmuel_2015	0.206 [0.010, 0.388]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.734 [0.605, 0.825]	100
SPG1_STRSG_Olson_2014	0.196 [−0.020, 0.397]	100

0.316

[0.254, 0.369]

1000

10 / 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	0.712 [0.610, 0.810]	100
CASP3_HUMAN_Roychowdhury_2020	0.703 [0.592, 0.806]	100
CCR5_HUMAN_Gill_2023	0.635 [0.528, 0.746]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.747 [0.640, 0.833]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.750 [0.632, 0.852]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.593 [0.470, 0.709]	100
MLAC_ECOLI_MacRae_2023	0.544 [0.428, 0.659]	100
MTH3_HAEAE_RockahShmuel_2015	0.651 [0.508, 0.781]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.932 [0.880, 0.976]	100
SPG1_STRSG_Olson_2014	0.762 [0.618, 0.886]	100

0.703

[0.667, 0.739]

1000

10 / 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	0.260 [0.066, 0.480]	100
CASP3_HUMAN_Roychowdhury_2020	0.341 [0.098, 0.510]	100
CCR5_HUMAN_Gill_2023	0.200 [0.020, 0.422]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.281 [0.100, 0.479]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.349 [0.145, 0.504]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.089 [−0.119, 0.281]	100
MLAC_ECOLI_MacRae_2023	0.080 [−0.160, 0.280]	100
MTH3_HAEAE_RockahShmuel_2015	0.094 [−0.074, 0.312]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.653 [0.540, 0.821]	100
SPG1_STRSG_Olson_2014	0.231 [0.043, 0.404]	100

0.258

[0.206, 0.327]

1000

10 / 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	0.785 [0.653, 0.917]	100
CASP3_HUMAN_Roychowdhury_2020	0.697 [0.639, 0.781]	100
CCR5_HUMAN_Gill_2023	0.806 [0.594, 0.875]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.813 [0.690, 0.898]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.828 [0.746, 0.907]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.582 [0.418, 0.709]	100
MLAC_ECOLI_MacRae_2023	0.772 [0.552, 0.904]	100
MTH3_HAEAE_RockahShmuel_2015	0.406 [0.212, 0.582]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.891 [0.806, 0.964]	100
SPG1_STRSG_Olson_2014	0.873 [0.724, 0.950]	100

0.745

[0.695, 0.779]

1000

10 / 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	0.100 [0.000, 0.500]	100
CASP3_HUMAN_Roychowdhury_2020	0.000 [0.000, 0.201]	100
CCR5_HUMAN_Gill_2023	0.100 [0.000, 0.400]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.300 [0.091, 0.600]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.100 [0.000, 0.364]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.200 [0.000, 0.500]	100
MLAC_ECOLI_MacRae_2023	0.200 [0.000, 0.500]	100
MTH3_HAEAE_RockahShmuel_2015	0.000 [0.000, 0.101]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.500 [0.100, 0.727]	100
SPG1_STRSG_Olson_2014	0.200 [0.000, 0.500]	100

0.170

[0.098, 0.257]

1000

4 / 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	0.578 [0.405, 0.709]	100
CASP3_HUMAN_Roychowdhury_2020	0.122 [−0.089, 0.333]	100
CCR5_HUMAN_Gill_2023	0.250 [0.064, 0.419]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.390 [0.195, 0.556]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.668 [0.522, 0.789]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.455 [0.264, 0.607]	100
MLAC_ECOLI_MacRae_2023	0.316 [0.113, 0.485]	100
MTH3_HAEAE_RockahShmuel_2015	0.309 [0.116, 0.471]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.890 [0.839, 0.921]	100
SPG1_STRSG_Olson_2014	0.784 [0.666, 0.863]	100

0.476

[0.418, 0.523]

1000

16 / 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	0.800 [0.713, 0.884]	100
CASP3_HUMAN_Roychowdhury_2020	0.531 [0.413, 0.647]	100
CCR5_HUMAN_Gill_2023	0.621 [0.513, 0.719]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.675 [0.556, 0.772]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.874 [0.783, 0.950]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.762 [0.641, 0.875]	100
MLAC_ECOLI_MacRae_2023	0.660 [0.552, 0.762]	100
MTH3_HAEAE_RockahShmuel_2015	0.672 [0.551, 0.789]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.956 [0.918, 0.989]	100
SPG1_STRSG_Olson_2014	0.941 [0.889, 0.978]	100

0.749

[0.718, 0.781]

1000

16 / 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	0.420 [0.221, 0.600]	100
CASP3_HUMAN_Roychowdhury_2020	0.060 [−0.160, 0.260]	100
CCR5_HUMAN_Gill_2023	0.080 [−0.081, 0.338]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.281 [0.043, 0.484]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.393 [0.193, 0.579]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.356 [0.178, 0.531]	100
MLAC_ECOLI_MacRae_2023	0.280 [0.041, 0.461]	100
MTH3_HAEAE_RockahShmuel_2015	0.234 [−0.013, 0.381]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.694 [0.555, 0.841]	100
SPG1_STRSG_Olson_2014	0.403 [0.295, 0.521]	100

0.320

[0.257, 0.378]

1000

16 / 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	0.784 [0.680, 0.921]	100
CASP3_HUMAN_Roychowdhury_2020	0.629 [0.389, 0.844]	100
CCR5_HUMAN_Gill_2023	0.808 [0.608, 0.879]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.588 [0.461, 0.718]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.924 [0.877, 0.965]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.652 [0.453, 0.812]	100
MLAC_ECOLI_MacRae_2023	0.856 [0.760, 0.934]	100
MTH3_HAEAE_RockahShmuel_2015	0.657 [0.439, 0.814]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.935 [0.906, 0.983]	100
SPG1_STRSG_Olson_2014	0.920 [0.850, 0.971]	100

0.775

[0.727, 0.816]

1000

15 / 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	0.400 [0.000, 0.546]	100
CASP3_HUMAN_Roychowdhury_2020	0.200 [0.000, 0.500]	100
CCR5_HUMAN_Gill_2023	0.100 [0.000, 0.455]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.000 [0.000, 0.100]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.200 [0.000, 0.600]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.100 [0.000, 0.364]	100
MLAC_ECOLI_MacRae_2023	0.100 [0.000, 0.500]	100
MTH3_HAEAE_RockahShmuel_2015	0.091 [0.000, 0.400]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.500 [0.200, 0.900]	100
SPG1_STRSG_Olson_2014	0.500 [0.182, 0.701]	100

0.219

[0.153, 0.318]

1000

12 / 17

0.360

reproduced

overlap

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

Board

tape protein-ml/tape

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	0.741	[0.724, 0.756]	100	10 / 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	0.585	[0.563, 0.607]	100	10 / 17	0.663	reproduced	overlap Train/test overlap declared by the evaluator: protein-ml/uniref50	Board

UniProt Knowledgebase protein-ml/uniprot-knowledgebase

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	0.684	[0.609, 0.779]	100	7 / 17	0.844	reproduced	unknown	Board
Fmax (biological process, probe) ↑ higher is better · probe	0.441	[0.352, 0.563]	100	12 / 17	0.538	reproduced	unknown	Board
Fmax (cellular component, probe) ↑ higher is better · probe	0.699	[0.644, 0.763]	100	17 / 17	0.756	reproduced	unknown	Board
Micro AUPR (molecular function, probe) ↑ higher is better · probe · secondary	0.729	[0.635, 0.820]	100	—	0.875	reproduced	unknown	Board
Smin (molecular function, probe) ↓ lower is better · bits · probe · secondary	3.206	[2.246, 4.008]	100	—	1.727	reproduced	unknown	Board
Micro AUPR (biological process, probe) ↑ higher is better · probe · secondary	0.467	[0.295, 0.609]	100	—	0.595	reproduced	unknown	Board
Smin (biological process, probe) ↓ lower is better · bits · probe · secondary	5.999	[4.547, 7.214]	100	—	4.842	reproduced	unknown	Board
Micro AUPR (cellular component, probe) ↑ higher is better · probe · secondary	0.673	[0.603, 0.741]	100	—	0.758	reproduced	unknown	Board
Smin (cellular component, probe) ↓ lower is better · bits · probe · secondary	3.621	[2.825, 4.407]	100	—	3.092	reproduced	unknown	Board

Other Benchmarks drawn from no hub dataset you can see

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	0.830	[0.813, 0.847]	1728	11 / 17	0.888	reproduced	overlap Train/test overlap declared by the evaluator: protein-ml/uniref50	Board
MCC (logistic probe) ↑ higher is better · probe · secondary	0.650	[0.615, 0.686]	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	0.256	[0.031, 0.446]	100	9 / 12	0.410	reproduced	unknown	Board
Spearman ρ (ridge probe) ↑ higher is better · probe	0.191	[−0.004, 0.366]	100	5 / 17	0.297	reproduced	unknown	Board
RMSE, kcal/mol (ridge probe) ↓ lower is better · kcal/mol · probe	1.109	[0.879, 1.303]	100	8 / 17	1.095	reproduced	unknown	Board
Pearson r (ridge probe) ↑ higher is better · probe · secondary	0.229	[0.060, 0.375]	100	—	0.349	reproduced	unknown	Board
MAE, kcal/mol (ridge probe) ↓ lower is better · kcal/mol · probe · secondary	0.859	[0.701, 1.021]	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	0.613	[0.597, 0.628]	2742	16 / 17	0.801	reproduced	overlap Train/test overlap declared by the evaluator: protein-ml/uniref50	Board
Macro F1 (logistic probe) ↑ higher is better · probe	0.416	[0.399, 0.433]	2742	16 / 17	0.640	reproduced	overlap Train/test overlap declared by the evaluator: protein-ml/uniref50	Board
MCC (logistic probe) ↑ higher is better · probe · secondary	0.526	[0.507, 0.545]	2742	—	0.758	reproduced	overlap Train/test overlap declared by the evaluator: protein-ml/uniref50	Board