Models / protein-ml / esm2-t30-150m-ur50d / benchmarks

protein-ml/esm2-t30-150m-ur50d

Visibility: PUBLIC

Model card FilesBenchmarks

Group by

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

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	0.278 [−0.132, 0.676]	25
MET_HUMAN_Estevam_2023	0.197 [−0.223, 0.588]	25
S22A1_HUMAN_Yee_2023_activity	0.349 [−0.042, 0.652]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.228 [−0.147, 0.581]	25

0.263

[0.046, 0.435]

100

9 / 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	0.654 [0.413, 0.860]	25
MET_HUMAN_Estevam_2023	0.682 [0.444, 0.885]	25
S22A1_HUMAN_Yee_2023_activity	0.613 [0.364, 0.840]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.701 [0.479, 0.894]	25

0.663

[0.548, 0.772]

100

7 / 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 [−0.282, 0.624]	25
MET_HUMAN_Estevam_2023	0.277 [−0.277, 0.686]	25
S22A1_HUMAN_Yee_2023_activity	0.131 [−0.275, 0.554]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.280 [−0.272, 0.714]	25

0.222

[−0.042, 0.425]

100

8 / 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	0.547 [0.216, 0.824]	25
MET_HUMAN_Estevam_2023	0.769 [0.296, 0.988]	25
S22A1_HUMAN_Yee_2023_activity	0.465 [0.094, 0.911]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.252 [0.059, 0.736]	25

0.508

[0.318, 0.707]

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, 0.000]	25
MET_HUMAN_Estevam_2023	0.000 [0.000, 0.667]	25
S22A1_HUMAN_Yee_2023_activity	0.000 [0.000, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.000 [0.000, 0.750]	25

0.000

[0.000, 0.375]

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	0.278 [−0.132, 0.676]	25
MET_HUMAN_Estevam_2023	0.197 [−0.223, 0.588]	25
S22A1_HUMAN_Yee_2023_activity	0.349 [−0.042, 0.652]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.228 [−0.147, 0.581]	25

0.263

[0.046, 0.435]

100

9 / 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	0.654 [0.413, 0.860]	25
MET_HUMAN_Estevam_2023	0.682 [0.444, 0.885]	25
S22A1_HUMAN_Yee_2023_activity	0.613 [0.364, 0.840]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.701 [0.479, 0.894]	25

0.663

[0.548, 0.772]

100

7 / 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 [−0.282, 0.624]	25
MET_HUMAN_Estevam_2023	0.277 [−0.277, 0.686]	25
S22A1_HUMAN_Yee_2023_activity	0.131 [−0.275, 0.554]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.280 [−0.272, 0.714]	25

0.222

[−0.042, 0.425]

100

8 / 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	0.547 [0.216, 0.824]	25
MET_HUMAN_Estevam_2023	0.769 [0.296, 0.988]	25
S22A1_HUMAN_Yee_2023_activity	0.465 [0.094, 0.911]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.252 [0.059, 0.736]	25

0.508

[0.318, 0.707]

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, 0.000]	25
MET_HUMAN_Estevam_2023	0.000 [0.000, 0.667]	25
S22A1_HUMAN_Yee_2023_activity	0.000 [0.000, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.000 [0.000, 0.750]	25

0.000

[0.000, 0.375]

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	0.118 [−0.320, 0.502]	25
CCR5_HUMAN_Gill_2023	0.278 [−0.132, 0.676]	25
OPSD_HUMAN_Wan_2019	0.412 [0.065, 0.702]	25
S22A1_HUMAN_Yee_2023_activity	0.349 [−0.042, 0.652]	25

0.289

[0.102, 0.483]

100

9 / 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	0.532 [0.292, 0.773]	25
CCR5_HUMAN_Gill_2023	0.654 [0.413, 0.860]	25
OPSD_HUMAN_Wan_2019	0.771 [0.573, 0.952]	25
S22A1_HUMAN_Yee_2023_activity	0.613 [0.364, 0.840]	25

0.643

[0.524, 0.764]

100

9 / 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 [−0.387, 0.471]	25
CCR5_HUMAN_Gill_2023	0.199 [−0.282, 0.624]	25
OPSD_HUMAN_Wan_2019	0.447 [0.007, 0.781]	25
S22A1_HUMAN_Yee_2023_activity	0.131 [−0.275, 0.554]	25

0.223

[−0.010, 0.424]

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
ADRB2_HUMAN_Jones_2020	0.636 [0.000, 1.000]	25
CCR5_HUMAN_Gill_2023	0.547 [0.216, 0.824]	25
OPSD_HUMAN_Wan_2019	0.278 [0.003, 0.743]	25
S22A1_HUMAN_Yee_2023_activity	0.465 [0.094, 0.911]	25

0.482

[0.269, 0.681]

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	0.333 [0.000, 1.000]	25
CCR5_HUMAN_Gill_2023	0.000 [0.000, 0.000]	25
OPSD_HUMAN_Wan_2019	0.000 [0.000, 0.715]	25
S22A1_HUMAN_Yee_2023_activity	0.000 [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	0.197 [−0.223, 0.588]	25
MK01_HUMAN_Brenan_2016	0.430 [0.003, 0.764]	25
SCN5A_HUMAN_Glazer_2019	0.188 [−0.276, 0.601]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.228 [−0.147, 0.581]	25

0.261

[0.040, 0.454]

100

6 / 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	0.682 [0.444, 0.885]	25
MK01_HUMAN_Brenan_2016	0.794 [0.515, 0.990]	25
SCN5A_HUMAN_Glazer_2019	0.533 [0.292, 0.753]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.701 [0.479, 0.894]	25

0.678

[0.561, 0.792]

100

5 / 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 [−0.277, 0.686]	25
MK01_HUMAN_Brenan_2016	0.316 [−0.022, 0.724]	25
SCN5A_HUMAN_Glazer_2019	−0.131 [−0.487, 0.368]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.280 [−0.272, 0.714]	25

0.186

[−0.024, 0.428]

100

8 / 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	0.769 [0.296, 0.988]	25
MK01_HUMAN_Brenan_2016	0.399 [0.000, 1.000]	25
SCN5A_HUMAN_Glazer_2019	0.986 [0.220, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.252 [0.059, 0.736]	25

0.601

[0.319, 0.807]

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, 0.667]	25
MK01_HUMAN_Brenan_2016	0.333 [0.000, 1.000]	25
SCN5A_HUMAN_Glazer_2019	0.667 [0.000, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.000 [0.000, 0.750]	25

0.250 best

[0.042, 0.542]

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

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	0.544 [0.394, 0.689]	100
CASP3_HUMAN_Roychowdhury_2020	0.625 [0.508, 0.725]	100
CCR5_HUMAN_Gill_2023	0.296 [0.115, 0.470]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.656 [0.517, 0.761]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.444 [0.248, 0.606]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.050 [−0.160, 0.241]	100
MLAC_ECOLI_MacRae_2023	0.247 [0.064, 0.413]	100
MTH3_HAEAE_RockahShmuel_2015	0.484 [0.307, 0.631]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.740 [0.610, 0.829]	100
SPG1_STRSG_Olson_2014	0.140 [−0.060, 0.344]	100

0.423

[0.361, 0.473]

1000

4 / 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	0.736 [0.642, 0.835]	100
CASP3_HUMAN_Roychowdhury_2020	0.844 [0.766, 0.920]	100
CCR5_HUMAN_Gill_2023	0.674 [0.564, 0.774]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.813 [0.717, 0.886]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.815 [0.721, 0.900]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.621 [0.488, 0.744]	100
MLAC_ECOLI_MacRae_2023	0.672 [0.565, 0.766]	100
MTH3_HAEAE_RockahShmuel_2015	0.787 [0.645, 0.899]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.929 [0.868, 0.976]	100
SPG1_STRSG_Olson_2014	0.795 [0.659, 0.908]	100

0.768

[0.735, 0.800]

1000

3 / 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 [0.098, 0.508]	100
CASP3_HUMAN_Roychowdhury_2020	0.541 [0.365, 0.712]	100
CCR5_HUMAN_Gill_2023	0.160 [0.000, 0.384]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.401 [0.240, 0.620]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.393 [0.192, 0.568]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.178 [−0.014, 0.383]	100
MLAC_ECOLI_MacRae_2023	0.320 [0.085, 0.490]	100
MTH3_HAEAE_RockahShmuel_2015	0.328 [0.115, 0.490]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.612 [0.449, 0.794]	100
SPG1_STRSG_Olson_2014	0.231 [0.081, 0.433]	100

0.342

[0.292, 0.410]

1000

6 / 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	0.783 [0.679, 0.915]	100
CASP3_HUMAN_Roychowdhury_2020	0.783 [0.729, 0.881]	100
CCR5_HUMAN_Gill_2023	0.843 [0.687, 0.884]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.856 [0.754, 0.939]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.847 [0.775, 0.925]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.492 [0.286, 0.708]	100
MLAC_ECOLI_MacRae_2023	0.836 [0.741, 0.902]	100
MTH3_HAEAE_RockahShmuel_2015	0.585 [0.330, 0.778]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.849 [0.740, 0.927]	100
SPG1_STRSG_Olson_2014	0.801 [0.671, 0.938]	100

0.768

[0.722, 0.809]

1000

8 / 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, 0.400]	100
CASP3_HUMAN_Roychowdhury_2020	0.100 [0.000, 0.500]	100
CCR5_HUMAN_Gill_2023	0.000 [0.000, 0.200]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.300 [0.083, 0.636]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.100 [0.000, 0.364]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.100 [0.000, 0.364]	100
MLAC_ECOLI_MacRae_2023	0.000 [0.000, 0.200]	100
MTH3_HAEAE_RockahShmuel_2015	0.091 [0.000, 0.364]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.300 [0.000, 0.600]	100
SPG1_STRSG_Olson_2014	0.200 [0.000, 0.500]	100

0.129

[0.078, 0.224]

1000

10 / 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	0.593 [0.427, 0.729]	100
CASP3_HUMAN_Roychowdhury_2020	0.593 [0.452, 0.709]	100
CCR5_HUMAN_Gill_2023	0.211 [0.007, 0.398]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.643 [0.494, 0.757]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.726 [0.573, 0.838]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.640 [0.473, 0.780]	100
MLAC_ECOLI_MacRae_2023	0.391 [0.182, 0.570]	100
MTH3_HAEAE_RockahShmuel_2015	0.520 [0.351, 0.656]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.884 [0.828, 0.917]	100
SPG1_STRSG_Olson_2014	0.840 [0.764, 0.891]	100

0.604

[0.551, 0.645]

1000

5 / 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	0.778 [0.688, 0.864]	100
CASP3_HUMAN_Roychowdhury_2020	0.809 [0.717, 0.893]	100
CCR5_HUMAN_Gill_2023	0.627 [0.514, 0.732]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.806 [0.711, 0.886]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.937 [0.871, 0.983]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.852 [0.736, 0.955]	100
MLAC_ECOLI_MacRae_2023	0.676 [0.571, 0.777]	100
MTH3_HAEAE_RockahShmuel_2015	0.774 [0.661, 0.877]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.968 [0.932, 0.994]	100
SPG1_STRSG_Olson_2014	0.940 [0.887, 0.980]	100

0.817

[0.788, 0.844]

1000

7 / 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	0.380 [0.179, 0.560]	100
CASP3_HUMAN_Roychowdhury_2020	0.501 [0.307, 0.681]	100
CCR5_HUMAN_Gill_2023	0.160 [−0.060, 0.343]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.441 [0.242, 0.636]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.611 [0.439, 0.743]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.490 [0.300, 0.633]	100
MLAC_ECOLI_MacRae_2023	0.280 [0.080, 0.469]	100
MTH3_HAEAE_RockahShmuel_2015	0.281 [0.134, 0.502]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.776 [0.603, 0.886]	100
SPG1_STRSG_Olson_2014	0.403 [0.295, 0.521]	100

0.432

[0.377, 0.491]

1000

9 / 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	0.833 [0.730, 0.965]	100
CASP3_HUMAN_Roychowdhury_2020	0.794 [0.671, 0.917]	100
CCR5_HUMAN_Gill_2023	0.779 [0.602, 0.847]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.823 [0.736, 0.918]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.952 [0.896, 0.979]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.573 [0.396, 0.788]	100
MLAC_ECOLI_MacRae_2023	0.846 [0.743, 0.929]	100
MTH3_HAEAE_RockahShmuel_2015	0.810 [0.620, 0.915]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.911 [0.850, 0.957]	100
SPG1_STRSG_Olson_2014	0.943 [0.902, 0.977]	100

0.827

[0.792, 0.865]

1000

8 / 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 [0.182, 0.727]	100
CASP3_HUMAN_Roychowdhury_2020	0.400 [0.091, 0.636]	100
CCR5_HUMAN_Gill_2023	0.000 [0.000, 0.300]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.500 [0.000, 0.700]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.300 [0.000, 0.600]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.200 [0.083, 0.600]	100
MLAC_ECOLI_MacRae_2023	0.200 [0.000, 0.500]	100
MTH3_HAEAE_RockahShmuel_2015	0.273 [0.000, 0.545]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.500 [0.100, 0.786]	100
SPG1_STRSG_Olson_2014	0.500 [0.200, 0.800]	100

0.327

[0.239, 0.421]

1000

3 / 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

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	0.278 [−0.132, 0.676]	25
MET_HUMAN_Estevam_2023	0.197 [−0.223, 0.588]	25
S22A1_HUMAN_Yee_2023_activity	0.349 [−0.042, 0.652]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.228 [−0.147, 0.581]	25

0.263

[0.046, 0.435]

100

9 / 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	0.654 [0.413, 0.860]	25
MET_HUMAN_Estevam_2023	0.682 [0.444, 0.885]	25
S22A1_HUMAN_Yee_2023_activity	0.613 [0.364, 0.840]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.701 [0.479, 0.894]	25

0.663

[0.548, 0.772]

100

7 / 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 [−0.282, 0.624]	25
MET_HUMAN_Estevam_2023	0.277 [−0.277, 0.686]	25
S22A1_HUMAN_Yee_2023_activity	0.131 [−0.275, 0.554]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.280 [−0.272, 0.714]	25

0.222

[−0.042, 0.425]

100

8 / 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	0.547 [0.216, 0.824]	25
MET_HUMAN_Estevam_2023	0.769 [0.296, 0.988]	25
S22A1_HUMAN_Yee_2023_activity	0.465 [0.094, 0.911]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.252 [0.059, 0.736]	25

0.508

[0.318, 0.707]

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, 0.000]	25
MET_HUMAN_Estevam_2023	0.000 [0.000, 0.667]	25
S22A1_HUMAN_Yee_2023_activity	0.000 [0.000, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.000 [0.000, 0.750]	25

0.000

[0.000, 0.375]

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	0.278 [−0.132, 0.676]	25
MET_HUMAN_Estevam_2023	0.197 [−0.223, 0.588]	25
S22A1_HUMAN_Yee_2023_activity	0.349 [−0.042, 0.652]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.228 [−0.147, 0.581]	25

0.263

[0.046, 0.435]

100

9 / 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	0.654 [0.413, 0.860]	25
MET_HUMAN_Estevam_2023	0.682 [0.444, 0.885]	25
S22A1_HUMAN_Yee_2023_activity	0.613 [0.364, 0.840]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.701 [0.479, 0.894]	25

0.663

[0.548, 0.772]

100

7 / 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 [−0.282, 0.624]	25
MET_HUMAN_Estevam_2023	0.277 [−0.277, 0.686]	25
S22A1_HUMAN_Yee_2023_activity	0.131 [−0.275, 0.554]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.280 [−0.272, 0.714]	25

0.222

[−0.042, 0.425]

100

8 / 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	0.547 [0.216, 0.824]	25
MET_HUMAN_Estevam_2023	0.769 [0.296, 0.988]	25
S22A1_HUMAN_Yee_2023_activity	0.465 [0.094, 0.911]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.252 [0.059, 0.736]	25

0.508

[0.318, 0.707]

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, 0.000]	25
MET_HUMAN_Estevam_2023	0.000 [0.000, 0.667]	25
S22A1_HUMAN_Yee_2023_activity	0.000 [0.000, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.000 [0.000, 0.750]	25

0.000

[0.000, 0.375]

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	0.118 [−0.320, 0.502]	25
CCR5_HUMAN_Gill_2023	0.278 [−0.132, 0.676]	25
OPSD_HUMAN_Wan_2019	0.412 [0.065, 0.702]	25
S22A1_HUMAN_Yee_2023_activity	0.349 [−0.042, 0.652]	25

0.289

[0.102, 0.483]

100

9 / 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	0.532 [0.292, 0.773]	25
CCR5_HUMAN_Gill_2023	0.654 [0.413, 0.860]	25
OPSD_HUMAN_Wan_2019	0.771 [0.573, 0.952]	25
S22A1_HUMAN_Yee_2023_activity	0.613 [0.364, 0.840]	25

0.643

[0.524, 0.764]

100

9 / 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 [−0.387, 0.471]	25
CCR5_HUMAN_Gill_2023	0.199 [−0.282, 0.624]	25
OPSD_HUMAN_Wan_2019	0.447 [0.007, 0.781]	25
S22A1_HUMAN_Yee_2023_activity	0.131 [−0.275, 0.554]	25

0.223

[−0.010, 0.424]

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
ADRB2_HUMAN_Jones_2020	0.636 [0.000, 1.000]	25
CCR5_HUMAN_Gill_2023	0.547 [0.216, 0.824]	25
OPSD_HUMAN_Wan_2019	0.278 [0.003, 0.743]	25
S22A1_HUMAN_Yee_2023_activity	0.465 [0.094, 0.911]	25

0.482

[0.269, 0.681]

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	0.333 [0.000, 1.000]	25
CCR5_HUMAN_Gill_2023	0.000 [0.000, 0.000]	25
OPSD_HUMAN_Wan_2019	0.000 [0.000, 0.715]	25
S22A1_HUMAN_Yee_2023_activity	0.000 [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	0.197 [−0.223, 0.588]	25
MK01_HUMAN_Brenan_2016	0.430 [0.003, 0.764]	25
SCN5A_HUMAN_Glazer_2019	0.188 [−0.276, 0.601]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.228 [−0.147, 0.581]	25

0.261

[0.040, 0.454]

100

6 / 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	0.682 [0.444, 0.885]	25
MK01_HUMAN_Brenan_2016	0.794 [0.515, 0.990]	25
SCN5A_HUMAN_Glazer_2019	0.533 [0.292, 0.753]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.701 [0.479, 0.894]	25

0.678

[0.561, 0.792]

100

5 / 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 [−0.277, 0.686]	25
MK01_HUMAN_Brenan_2016	0.316 [−0.022, 0.724]	25
SCN5A_HUMAN_Glazer_2019	−0.131 [−0.487, 0.368]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.280 [−0.272, 0.714]	25

0.186

[−0.024, 0.428]

100

8 / 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	0.769 [0.296, 0.988]	25
MK01_HUMAN_Brenan_2016	0.399 [0.000, 1.000]	25
SCN5A_HUMAN_Glazer_2019	0.986 [0.220, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.252 [0.059, 0.736]	25

0.601

[0.319, 0.807]

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, 0.667]	25
MK01_HUMAN_Brenan_2016	0.333 [0.000, 1.000]	25
SCN5A_HUMAN_Glazer_2019	0.667 [0.000, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.000 [0.000, 0.750]	25

0.250 best

[0.042, 0.542]

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

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	0.544 [0.394, 0.689]	100
CASP3_HUMAN_Roychowdhury_2020	0.625 [0.508, 0.725]	100
CCR5_HUMAN_Gill_2023	0.296 [0.115, 0.470]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.656 [0.517, 0.761]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.444 [0.248, 0.606]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.050 [−0.160, 0.241]	100
MLAC_ECOLI_MacRae_2023	0.247 [0.064, 0.413]	100
MTH3_HAEAE_RockahShmuel_2015	0.484 [0.307, 0.631]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.740 [0.610, 0.829]	100
SPG1_STRSG_Olson_2014	0.140 [−0.060, 0.344]	100

0.423

[0.361, 0.473]

1000

4 / 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	0.736 [0.642, 0.835]	100
CASP3_HUMAN_Roychowdhury_2020	0.844 [0.766, 0.920]	100
CCR5_HUMAN_Gill_2023	0.674 [0.564, 0.774]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.813 [0.717, 0.886]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.815 [0.721, 0.900]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.621 [0.488, 0.744]	100
MLAC_ECOLI_MacRae_2023	0.672 [0.565, 0.766]	100
MTH3_HAEAE_RockahShmuel_2015	0.787 [0.645, 0.899]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.929 [0.868, 0.976]	100
SPG1_STRSG_Olson_2014	0.795 [0.659, 0.908]	100

0.768

[0.735, 0.800]

1000

3 / 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 [0.098, 0.508]	100
CASP3_HUMAN_Roychowdhury_2020	0.541 [0.365, 0.712]	100
CCR5_HUMAN_Gill_2023	0.160 [0.000, 0.384]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.401 [0.240, 0.620]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.393 [0.192, 0.568]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.178 [−0.014, 0.383]	100
MLAC_ECOLI_MacRae_2023	0.320 [0.085, 0.490]	100
MTH3_HAEAE_RockahShmuel_2015	0.328 [0.115, 0.490]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.612 [0.449, 0.794]	100
SPG1_STRSG_Olson_2014	0.231 [0.081, 0.433]	100

0.342

[0.292, 0.410]

1000

6 / 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	0.783 [0.679, 0.915]	100
CASP3_HUMAN_Roychowdhury_2020	0.783 [0.729, 0.881]	100
CCR5_HUMAN_Gill_2023	0.843 [0.687, 0.884]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.856 [0.754, 0.939]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.847 [0.775, 0.925]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.492 [0.286, 0.708]	100
MLAC_ECOLI_MacRae_2023	0.836 [0.741, 0.902]	100
MTH3_HAEAE_RockahShmuel_2015	0.585 [0.330, 0.778]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.849 [0.740, 0.927]	100
SPG1_STRSG_Olson_2014	0.801 [0.671, 0.938]	100

0.768

[0.722, 0.809]

1000

8 / 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, 0.400]	100
CASP3_HUMAN_Roychowdhury_2020	0.100 [0.000, 0.500]	100
CCR5_HUMAN_Gill_2023	0.000 [0.000, 0.200]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.300 [0.083, 0.636]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.100 [0.000, 0.364]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.100 [0.000, 0.364]	100
MLAC_ECOLI_MacRae_2023	0.000 [0.000, 0.200]	100
MTH3_HAEAE_RockahShmuel_2015	0.091 [0.000, 0.364]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.300 [0.000, 0.600]	100
SPG1_STRSG_Olson_2014	0.200 [0.000, 0.500]	100

0.129

[0.078, 0.224]

1000

10 / 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	0.593 [0.427, 0.729]	100
CASP3_HUMAN_Roychowdhury_2020	0.593 [0.452, 0.709]	100
CCR5_HUMAN_Gill_2023	0.211 [0.007, 0.398]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.643 [0.494, 0.757]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.726 [0.573, 0.838]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.640 [0.473, 0.780]	100
MLAC_ECOLI_MacRae_2023	0.391 [0.182, 0.570]	100
MTH3_HAEAE_RockahShmuel_2015	0.520 [0.351, 0.656]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.884 [0.828, 0.917]	100
SPG1_STRSG_Olson_2014	0.840 [0.764, 0.891]	100

0.604

[0.551, 0.645]

1000

5 / 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	0.778 [0.688, 0.864]	100
CASP3_HUMAN_Roychowdhury_2020	0.809 [0.717, 0.893]	100
CCR5_HUMAN_Gill_2023	0.627 [0.514, 0.732]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.806 [0.711, 0.886]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.937 [0.871, 0.983]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.852 [0.736, 0.955]	100
MLAC_ECOLI_MacRae_2023	0.676 [0.571, 0.777]	100
MTH3_HAEAE_RockahShmuel_2015	0.774 [0.661, 0.877]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.968 [0.932, 0.994]	100
SPG1_STRSG_Olson_2014	0.940 [0.887, 0.980]	100

0.817

[0.788, 0.844]

1000

7 / 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	0.380 [0.179, 0.560]	100
CASP3_HUMAN_Roychowdhury_2020	0.501 [0.307, 0.681]	100
CCR5_HUMAN_Gill_2023	0.160 [−0.060, 0.343]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.441 [0.242, 0.636]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.611 [0.439, 0.743]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.490 [0.300, 0.633]	100
MLAC_ECOLI_MacRae_2023	0.280 [0.080, 0.469]	100
MTH3_HAEAE_RockahShmuel_2015	0.281 [0.134, 0.502]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.776 [0.603, 0.886]	100
SPG1_STRSG_Olson_2014	0.403 [0.295, 0.521]	100

0.432

[0.377, 0.491]

1000

9 / 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	0.833 [0.730, 0.965]	100
CASP3_HUMAN_Roychowdhury_2020	0.794 [0.671, 0.917]	100
CCR5_HUMAN_Gill_2023	0.779 [0.602, 0.847]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.823 [0.736, 0.918]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.952 [0.896, 0.979]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.573 [0.396, 0.788]	100
MLAC_ECOLI_MacRae_2023	0.846 [0.743, 0.929]	100
MTH3_HAEAE_RockahShmuel_2015	0.810 [0.620, 0.915]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.911 [0.850, 0.957]	100
SPG1_STRSG_Olson_2014	0.943 [0.902, 0.977]	100

0.827

[0.792, 0.865]

1000

8 / 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 [0.182, 0.727]	100
CASP3_HUMAN_Roychowdhury_2020	0.400 [0.091, 0.636]	100
CCR5_HUMAN_Gill_2023	0.000 [0.000, 0.300]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.500 [0.000, 0.700]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.300 [0.000, 0.600]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.200 [0.083, 0.600]	100
MLAC_ECOLI_MacRae_2023	0.200 [0.000, 0.500]	100
MTH3_HAEAE_RockahShmuel_2015	0.273 [0.000, 0.545]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.500 [0.100, 0.786]	100
SPG1_STRSG_Olson_2014	0.500 [0.200, 0.800]	100

0.327

[0.239, 0.421]

1000

3 / 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	0.803	[0.787, 0.818]	100	3 / 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	0.650	[0.631, 0.669]	100	4 / 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	0.652	[0.590, 0.711]	100	12 / 17	0.844	reproduced	unknown	Board
Fmax (biological process, probe) ↑ higher is better · probe	0.488	[0.381, 0.596]	100	5 / 17	0.538	reproduced	unknown	Board
Fmax (cellular component, probe) ↑ higher is better · probe	0.722	[0.671, 0.784]	100	13 / 17	0.756	reproduced	unknown	Board
Micro AUPR (molecular function, probe) ↑ higher is better · probe · secondary	0.637	[0.547, 0.736]	100	—	0.875	reproduced	unknown	Board
Smin (molecular function, probe) ↓ lower is better · bits · probe · secondary	3.702	[2.859, 4.331]	100	—	1.727	reproduced	unknown	Board
Micro AUPR (biological process, probe) ↑ higher is better · probe · secondary	0.399	[0.255, 0.556]	100	—	0.595	reproduced	unknown	Board
Smin (biological process, probe) ↓ lower is better · bits · probe · secondary	6.062	[4.238, 7.687]	100	—	4.842	reproduced	unknown	Board
Micro AUPR (cellular component, probe) ↑ higher is better · probe · secondary	0.716	[0.653, 0.780]	100	—	0.758	reproduced	unknown	Board
Smin (cellular component, probe) ↓ lower is better · bits · probe · secondary	3.460	[2.662, 4.100]	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	0.855	[0.839, 0.872]	1728	5 / 17	0.888	reproduced	overlap Train/test overlap declared by the evaluator: protein-ml/uniref50	Board
MCC (logistic probe) ↑ higher is better · probe · secondary	0.708	[0.676, 0.741]	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	0.247	[0.023, 0.465]	100	10 / 12	0.410	reproduced	unknown	Board
Spearman ρ (ridge probe) ↑ higher is better · probe	0.094	[−0.090, 0.275]	100	13 / 17	0.297	reproduced	unknown	Board
RMSE, kcal/mol (ridge probe) ↓ lower is better · kcal/mol · probe	1.139	[0.910, 1.332]	100	15 / 17	1.095	reproduced	unknown	Board
Pearson r (ridge probe) ↑ higher is better · probe · secondary	0.132	[−0.071, 0.270]	100	—	0.349	reproduced	unknown	Board
MAE, kcal/mol (ridge probe) ↓ lower is better · kcal/mol · probe · secondary	0.859	[0.708, 1.013]	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	0.698	[0.683, 0.712]	2742	7 / 17	0.801	reproduced	overlap Train/test overlap declared by the evaluator: protein-ml/uniref50	Board
Macro F1 (logistic probe) ↑ higher is better · probe	0.510	[0.492, 0.528]	2742	7 / 17	0.640	reproduced	overlap Train/test overlap declared by the evaluator: protein-ml/uniref50	Board
MCC (logistic probe) ↑ higher is better · probe · secondary	0.633	[0.615, 0.651]	2742	—	0.758	reproduced	overlap Train/test overlap declared by the evaluator: protein-ml/uniref50	Board