Models / protein-ml / amplify-120m / benchmarks

protein-ml/amplify-120m

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	0.391 [−0.039, 0.711]	25
MET_HUMAN_Estevam_2023	0.322 [−0.061, 0.653]	25
S22A1_HUMAN_Yee_2023_activity	0.464 [0.094, 0.709]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.229 [−0.121, 0.521]	25

0.351

[0.150, 0.501]

100

4 / 13

0.476

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

Board

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

By assay: 4 levels

assay	Value [95% CI]	n
CCR5_HUMAN_Gill_2023	0.673 [0.436, 0.903]	25
MET_HUMAN_Estevam_2023	0.766 [0.561, 0.941]	25
S22A1_HUMAN_Yee_2023_activity	0.773 [0.570, 0.929]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.688 [0.441, 0.893]	25

0.725

[0.611, 0.821]

100

2 / 13

0.759

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

Board

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

By assay: 4 levels

assay	Value [95% CI]	n
CCR5_HUMAN_Gill_2023	0.359 [−0.099, 0.761]	25
MET_HUMAN_Estevam_2023	0.439 [−0.039, 0.763]	25
S22A1_HUMAN_Yee_2023_activity	0.458 [−0.026, 0.757]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.113 [−0.243, 0.601]	25

0.342

[0.108, 0.537]

100

4 / 13

0.426

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

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	0.185 [0.000, 0.830]	25
MET_HUMAN_Estevam_2023	0.877 [0.737, 0.984]	25
S22A1_HUMAN_Yee_2023_activity	0.813 [0.662, 0.916]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.256 [0.070, 0.698]	25

0.533

[0.446, 0.739]

100

—

0.635

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

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, 0.667]	25
MET_HUMAN_Estevam_2023	0.000 [0.000, 0.750]	25
S22A1_HUMAN_Yee_2023_activity	0.000 [0.000, 0.805]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.000 [0.000, 0.667]	25

0.000

[0.000, 0.396]

100

—

0.167

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

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	0.391 [−0.039, 0.711]	25
MET_HUMAN_Estevam_2023	0.322 [−0.061, 0.653]	25
S22A1_HUMAN_Yee_2023_activity	0.464 [0.094, 0.709]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.229 [−0.121, 0.521]	25

0.351

[0.150, 0.501]

100

4 / 13

0.476

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

Board

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

By assay: 4 levels

assay	Value [95% CI]	n
CCR5_HUMAN_Gill_2023	0.673 [0.436, 0.903]	25
MET_HUMAN_Estevam_2023	0.766 [0.561, 0.941]	25
S22A1_HUMAN_Yee_2023_activity	0.773 [0.570, 0.929]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.688 [0.441, 0.893]	25

0.725

[0.611, 0.821]

100

2 / 13

0.759

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

Board

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

By assay: 4 levels

assay	Value [95% CI]	n
CCR5_HUMAN_Gill_2023	0.359 [−0.099, 0.761]	25
MET_HUMAN_Estevam_2023	0.439 [−0.039, 0.763]	25
S22A1_HUMAN_Yee_2023_activity	0.458 [−0.026, 0.757]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.113 [−0.243, 0.601]	25

0.342

[0.108, 0.537]

100

4 / 13

0.426

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

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	0.185 [0.000, 0.830]	25
MET_HUMAN_Estevam_2023	0.877 [0.737, 0.984]	25
S22A1_HUMAN_Yee_2023_activity	0.813 [0.662, 0.916]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.256 [0.070, 0.698]	25

0.533

[0.446, 0.739]

100

—

0.635

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

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, 0.667]	25
MET_HUMAN_Estevam_2023	0.000 [0.000, 0.750]	25
S22A1_HUMAN_Yee_2023_activity	0.000 [0.000, 0.805]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.000 [0.000, 0.667]	25

0.000

[0.000, 0.396]

100

—

0.167

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

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	0.220 [−0.241, 0.674]	25
CCR5_HUMAN_Gill_2023	0.391 [−0.039, 0.711]	25
OPSD_HUMAN_Wan_2019	0.395 [−0.071, 0.733]	25
S22A1_HUMAN_Yee_2023_activity	0.464 [0.094, 0.709]	25

0.367

[0.141, 0.541]

100

4 / 13

0.463

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

Board

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

By assay: 4 levels

assay	Value [95% CI]	n
ADRB2_HUMAN_Jones_2020	0.588 [0.313, 0.842]	25
CCR5_HUMAN_Gill_2023	0.673 [0.436, 0.903]	25
OPSD_HUMAN_Wan_2019	0.694 [0.433, 0.929]	25
S22A1_HUMAN_Yee_2023_activity	0.773 [0.570, 0.929]	25

0.682

[0.553, 0.796]

100

5 / 13

0.751

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

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 [−0.316, 0.560]	25
CCR5_HUMAN_Gill_2023	0.359 [−0.099, 0.761]	25
OPSD_HUMAN_Wan_2019	0.113 [−0.263, 0.688]	25
S22A1_HUMAN_Yee_2023_activity	0.458 [−0.026, 0.757]	25

0.262

[0.038, 0.481]

100

6 / 13

0.426

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

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	0.657 [0.463, 0.976]	25
CCR5_HUMAN_Gill_2023	0.185 [0.000, 0.830]	25
OPSD_HUMAN_Wan_2019	0.920 [0.629, 1.000]	25
S22A1_HUMAN_Yee_2023_activity	0.813 [0.662, 0.916]	25

0.644

[0.523, 0.828]

100

—

0.709

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

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, 1.000]	25
CCR5_HUMAN_Gill_2023	0.000 [0.000, 0.667]	25
OPSD_HUMAN_Wan_2019	0.333 [0.000, 1.000]	25
S22A1_HUMAN_Yee_2023_activity	0.000 [0.000, 0.805]	25

0.083

[0.000, 0.500]

100

—

0.250

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

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	0.322 [−0.061, 0.653]	25
MK01_HUMAN_Brenan_2016	0.031 [−0.375, 0.493]	25
SCN5A_HUMAN_Glazer_2019	0.174 [−0.292, 0.578]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.229 [−0.121, 0.521]	25

0.189

[−0.028, 0.377]

100

10 / 13

0.362

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

Board

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

By assay: 4 levels

assay	Value [95% CI]	n
MET_HUMAN_Estevam_2023	0.766 [0.561, 0.941]	25
MK01_HUMAN_Brenan_2016	0.603 [0.325, 0.867]	25
SCN5A_HUMAN_Glazer_2019	0.553 [0.302, 0.786]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.688 [0.441, 0.893]	25

0.653

[0.534, 0.763]

100

10 / 13

0.722

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

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 [−0.039, 0.763]	25
MK01_HUMAN_Brenan_2016	0.144 [−0.342, 0.553]	25
SCN5A_HUMAN_Glazer_2019	0.033 [−0.458, 0.447]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.113 [−0.243, 0.601]	25

0.182

[−0.059, 0.381]

100

9 / 13

0.310

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

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 [0.737, 0.984]	25
MK01_HUMAN_Brenan_2016	0.589 [0.000, 0.956]	25
SCN5A_HUMAN_Glazer_2019	0.800 [0.160, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.256 [0.070, 0.698]	25

0.631

[0.412, 0.823]

100

—

0.725

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

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

0.083

[0.000, 0.500]

100

—

0.250

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

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	0.651	[0.583, 0.738]	100	13 / 17	0.844	reproduced	unknown	Board
Fmax (biological process, probe) ↑ higher is better · probe	0.456	[0.370, 0.576]	100	9 / 17	0.538	reproduced	unknown	Board
Fmax (cellular component, probe) ↑ higher is better · probe	0.723	[0.671, 0.780]	100	12 / 17	0.756	reproduced	unknown	Board
Micro AUPR (molecular function, probe) ↑ higher is better · probe · secondary	0.652	[0.558, 0.753]	100	—	0.875	reproduced	unknown	Board
Smin (molecular function, probe) ↓ lower is better · bits · probe · secondary	3.569	[2.652, 4.354]	100	—	1.727	reproduced	unknown	Board
Micro AUPR (biological process, probe) ↑ higher is better · probe · secondary	0.462	[0.317, 0.591]	100	—	0.595	reproduced	unknown	Board
Smin (biological process, probe) ↓ lower is better · bits · probe · secondary	5.807	[4.142, 7.443]	100	—	4.842	reproduced	unknown	Board
Micro AUPR (cellular component, probe) ↑ higher is better · probe · secondary	0.713	[0.653, 0.772]	100	—	0.758	reproduced	unknown	Board
Smin (cellular component, probe) ↓ lower is better · bits · probe · secondary	3.418	[2.639, 4.167]	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	0.822	[0.806, 0.841]	1728	15 / 17	0.888	reproduced	overlap Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)	Board
MCC (logistic probe) ↑ higher is better · probe · secondary	0.637	[0.601, 0.677]	1728	—	0.771	reproduced	overlap Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)	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	0.296	[0.098, 0.468]	100	7 / 12	0.410	reproduced	unknown	Board
Spearman ρ (ridge probe) ↑ higher is better · probe	0.083	[−0.148, 0.288]	100	14 / 17	0.297	reproduced	unknown	Board
RMSE, kcal/mol (ridge probe) ↓ lower is better · kcal/mol · probe	1.103	[0.875, 1.306]	100	5 / 17	1.095	reproduced	unknown	Board
Pearson r (ridge probe) ↑ higher is better · probe · secondary	0.141	[−0.053, 0.303]	100	—	0.349	reproduced	unknown	Board
MAE, kcal/mol (ridge probe) ↓ lower is better · kcal/mol · probe · secondary	0.845	[0.702, 1.001]	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	0.637	[0.622, 0.650]	2742	14 / 17	0.801	reproduced	overlap Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)	Board
Macro F1 (logistic probe) ↑ higher is better · probe	0.457	[0.438, 0.476]	2742	14 / 17	0.640	reproduced	overlap Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)	Board
MCC (logistic probe) ↑ higher is better · probe · secondary	0.559	[0.540, 0.576]	2742	—	0.758	reproduced	overlap Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)	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	0.469 [0.310, 0.624]	100
CASP3_HUMAN_Roychowdhury_2020	0.553 [0.399, 0.679]	100
CCR5_HUMAN_Gill_2023	0.284 [0.099, 0.463]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.567 [0.382, 0.702]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.300 [0.106, 0.479]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.090 [−0.123, 0.279]	100
MLAC_ECOLI_MacRae_2023	0.402 [0.223, 0.556]	100
MTH3_HAEAE_RockahShmuel_2015	0.696 [0.582, 0.781]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.794 [0.673, 0.872]	100
SPG1_STRSG_Olson_2014	0.184 [−0.018, 0.384]	100

0.434 best

[0.376, 0.483]

1000

1 / 11

0.434

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

Board

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

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.679 [0.576, 0.788]	100
CASP3_HUMAN_Roychowdhury_2020	0.808 [0.720, 0.890]	100
CCR5_HUMAN_Gill_2023	0.627 [0.520, 0.731]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.762 [0.656, 0.845]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.666 [0.540, 0.787]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.669 [0.555, 0.784]	100
MLAC_ECOLI_MacRae_2023	0.770 [0.669, 0.859]	100
MTH3_HAEAE_RockahShmuel_2015	0.853 [0.770, 0.928]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.961 [0.922, 0.990]	100
SPG1_STRSG_Olson_2014	0.797 [0.673, 0.899]	100

0.759

[0.727, 0.788]

1000

4 / 11

0.775

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

Board

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

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.380 [0.163, 0.540]	100
CASP3_HUMAN_Roychowdhury_2020	0.381 [0.201, 0.580]	100
CCR5_HUMAN_Gill_2023	0.240 [0.020, 0.403]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.281 [0.098, 0.500]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.175 [−0.026, 0.386]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.267 [0.074, 0.447]	100
MLAC_ECOLI_MacRae_2023	0.360 [0.179, 0.569]	100
MTH3_HAEAE_RockahShmuel_2015	0.421 [0.290, 0.598]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.694 [0.563, 0.880]	100
SPG1_STRSG_Olson_2014	0.346 [0.193, 0.469]	100

0.355

[0.301, 0.417]

1000

3 / 11

0.370

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

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 [0.676, 0.923]	100
CASP3_HUMAN_Roychowdhury_2020	0.807 [0.712, 0.891]	100
CCR5_HUMAN_Gill_2023	0.863 [0.629, 0.927]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.866 [0.802, 0.927]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.811 [0.720, 0.892]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.588 [0.441, 0.745]	100
MLAC_ECOLI_MacRae_2023	0.883 [0.792, 0.940]	100
MTH3_HAEAE_RockahShmuel_2015	0.795 [0.708, 0.881]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.854 [0.779, 0.936]	100
SPG1_STRSG_Olson_2014	0.876 [0.763, 0.943]	100

0.813

[0.775, 0.839]

1000

3 / 11

0.814

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

Board

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

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.200 [0.000, 0.463]	100
CASP3_HUMAN_Roychowdhury_2020	0.100 [0.000, 0.300]	100
CCR5_HUMAN_Gill_2023	0.300 [0.000, 0.545]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.300 [0.091, 0.700]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.000 [0.000, 0.300]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.200 [0.000, 0.455]	100
MLAC_ECOLI_MacRae_2023	0.100 [0.000, 0.333]	100
MTH3_HAEAE_RockahShmuel_2015	0.182 [0.000, 0.500]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.300 [0.000, 0.600]	100
SPG1_STRSG_Olson_2014	0.200 [0.000, 0.500]	100

0.188

[0.110, 0.278]

1000

2 / 11

0.196

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

Board

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

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.588 [0.427, 0.707]	100
CASP3_HUMAN_Roychowdhury_2020	0.528 [0.370, 0.656]	100
CCR5_HUMAN_Gill_2023	0.271 [0.100, 0.444]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.607 [0.440, 0.726]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.631 [0.453, 0.768]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.715 [0.594, 0.812]	100
MLAC_ECOLI_MacRae_2023	0.391 [0.185, 0.560]	100
MTH3_HAEAE_RockahShmuel_2015	0.557 [0.414, 0.681]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.879 [0.811, 0.918]	100
SPG1_STRSG_Olson_2014	0.846 [0.764, 0.899]	100

0.601

[0.554, 0.642]

1000

6 / 17

0.661

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

Board

AUC (ridge probe) ↑ higher is better · probe

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.779 [0.689, 0.864]	100
CASP3_HUMAN_Roychowdhury_2020	0.796 [0.701, 0.877]	100
CCR5_HUMAN_Gill_2023	0.624 [0.515, 0.734]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.778 [0.676, 0.861]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.863 [0.769, 0.937]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.860 [0.761, 0.944]	100
MLAC_ECOLI_MacRae_2023	0.690 [0.580, 0.794]	100
MTH3_HAEAE_RockahShmuel_2015	0.805 [0.695, 0.895]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.962 [0.922, 0.992]	100
SPG1_STRSG_Olson_2014	0.944 [0.896, 0.982]	100

0.810

[0.781, 0.838]

1000

8 / 17

0.848

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

Board

MCC (ridge probe) ↑ higher is better · probe

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.460 [0.264, 0.620]	100
CASP3_HUMAN_Roychowdhury_2020	0.461 [0.266, 0.640]	100
CCR5_HUMAN_Gill_2023	0.240 [0.000, 0.420]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.361 [0.141, 0.545]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.480 [0.305, 0.642]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.490 [0.307, 0.650]	100
MLAC_ECOLI_MacRae_2023	0.360 [0.156, 0.541]	100
MTH3_HAEAE_RockahShmuel_2015	0.328 [0.140, 0.490]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.776 [0.581, 0.864]	100
SPG1_STRSG_Olson_2014	0.403 [0.289, 0.521]	100

0.436

[0.369, 0.480]

1000

8 / 17

0.472

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

Board

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

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.801 [0.680, 0.932]	100
CASP3_HUMAN_Roychowdhury_2020	0.800 [0.613, 0.912]	100
CCR5_HUMAN_Gill_2023	0.818 [0.629, 0.875]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.848 [0.716, 0.936]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.940 [0.889, 0.979]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.728 [0.550, 0.878]	100
MLAC_ECOLI_MacRae_2023	0.869 [0.767, 0.930]	100
MTH3_HAEAE_RockahShmuel_2015	0.750 [0.593, 0.862]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.921 [0.882, 0.974]	100
SPG1_STRSG_Olson_2014	0.933 [0.913, 0.974]	100

0.841

[0.799, 0.870]

1000

3 / 17

0.860

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

Board

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

By assay: 10 levels

assay	Value [95% CI]	n
ANCSZ_Hobbs_2022	0.300 [0.000, 0.501]	100
CASP3_HUMAN_Roychowdhury_2020	0.300 [0.000, 0.545]	100
CCR5_HUMAN_Gill_2023	0.100 [0.000, 0.364]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.500 [0.100, 0.700]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.300 [0.083, 0.700]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.400 [0.100, 0.700]	100
MLAC_ECOLI_MacRae_2023	0.200 [0.000, 0.500]	100
MTH3_HAEAE_RockahShmuel_2015	0.091 [0.000, 0.400]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.500 [0.167, 0.800]	100
SPG1_STRSG_Olson_2014	0.300 [0.091, 0.700]	100

0.299

[0.201, 0.386]

1000

5 / 17

0.360

reproduced

overlap

Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)

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	0.751	[0.735, 0.768]	100	8 / 17	0.813	reproduced	overlap Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)	Board
Q8 accuracy (per-residue probe) ↑ higher is better · probe	0.597	[0.577, 0.620]	100	8 / 17	0.663	reproduced	overlap Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)	Board