Models / protein-ml / amplify-350m / benchmarks

protein-ml/amplify-350m

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	0.205 [−0.242, 0.582]	25
MET_HUMAN_Estevam_2023	0.156 [−0.277, 0.562]	25
S22A1_HUMAN_Yee_2023_activity	0.482 [0.081, 0.753]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	−0.085 [−0.453, 0.292]	25

0.189

[−0.026, 0.378]

100

12 / 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	0.590 [0.333, 0.808]	25
MET_HUMAN_Estevam_2023	0.669 [0.429, 0.883]	25
S22A1_HUMAN_Yee_2023_activity	0.700 [0.436, 0.926]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.507 [0.260, 0.753]	25

0.616

[0.489, 0.738]

100

12 / 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	0.038 [−0.428, 0.445]	25
MET_HUMAN_Estevam_2023	0.116 [−0.294, 0.650]	25
S22A1_HUMAN_Yee_2023_activity	0.294 [−0.175, 0.690]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	−0.053 [−0.442, 0.368]	25

0.099

[−0.120, 0.324]

100

13 / 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	0.555 [0.000, 0.798]	25
MET_HUMAN_Estevam_2023	0.901 [0.720, 0.992]	25
S22A1_HUMAN_Yee_2023_activity	0.653 [0.101, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.250 [0.039, 0.593]	25

0.590

[0.412, 0.747]

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, 0.500]	25
MET_HUMAN_Estevam_2023	0.333 [0.000, 1.000]	25
S22A1_HUMAN_Yee_2023_activity	0.333 [0.000, 0.800]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.000 [0.000, 0.000]	25

0.167 best

[0.000, 0.383]

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	0.205 [−0.242, 0.582]	25
MET_HUMAN_Estevam_2023	0.156 [−0.277, 0.562]	25
S22A1_HUMAN_Yee_2023_activity	0.482 [0.081, 0.753]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	−0.085 [−0.453, 0.292]	25

0.189

[−0.026, 0.378]

100

12 / 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	0.590 [0.333, 0.808]	25
MET_HUMAN_Estevam_2023	0.669 [0.429, 0.883]	25
S22A1_HUMAN_Yee_2023_activity	0.700 [0.436, 0.926]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.507 [0.260, 0.753]	25

0.616

[0.489, 0.738]

100

13 / 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	0.038 [−0.428, 0.445]	25
MET_HUMAN_Estevam_2023	0.116 [−0.294, 0.650]	25
S22A1_HUMAN_Yee_2023_activity	0.294 [−0.175, 0.690]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	−0.053 [−0.442, 0.368]	25

0.099

[−0.120, 0.324]

100

13 / 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	0.555 [0.000, 0.798]	25
MET_HUMAN_Estevam_2023	0.901 [0.720, 0.992]	25
S22A1_HUMAN_Yee_2023_activity	0.653 [0.101, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.250 [0.039, 0.593]	25

0.590

[0.412, 0.747]

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, 0.500]	25
MET_HUMAN_Estevam_2023	0.333 [0.000, 1.000]	25
S22A1_HUMAN_Yee_2023_activity	0.333 [0.000, 0.800]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.000 [0.000, 0.000]	25

0.167 best

[0.000, 0.383]

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	0.332 [−0.103, 0.672]	25
CCR5_HUMAN_Gill_2023	0.205 [−0.242, 0.582]	25
OPSD_HUMAN_Wan_2019	0.303 [−0.152, 0.670]	25
S22A1_HUMAN_Yee_2023_activity	0.482 [0.081, 0.753]	25

0.330

[0.119, 0.512]

100

7 / 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	0.649 [0.400, 0.868]	25
CCR5_HUMAN_Gill_2023	0.590 [0.333, 0.808]	25
OPSD_HUMAN_Wan_2019	0.674 [0.381, 0.939]	25
S22A1_HUMAN_Yee_2023_activity	0.700 [0.436, 0.926]	25

0.653

[0.520, 0.767]

100

7 / 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	0.277 [−0.144, 0.679]	25
CCR5_HUMAN_Gill_2023	0.038 [−0.428, 0.445]	25
OPSD_HUMAN_Wan_2019	0.280 [−0.167, 0.704]	25
S22A1_HUMAN_Yee_2023_activity	0.294 [−0.175, 0.690]	25

0.223

[0.008, 0.417]

100

8 / 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	0.438 [0.285, 0.986]	25
CCR5_HUMAN_Gill_2023	0.555 [0.000, 0.798]	25
OPSD_HUMAN_Wan_2019	0.685 [0.423, 1.000]	25
S22A1_HUMAN_Yee_2023_activity	0.653 [0.101, 1.000]	25

0.583

[0.403, 0.790]

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	0.333 [0.000, 1.000]	25
CCR5_HUMAN_Gill_2023	0.000 [0.000, 0.500]	25
OPSD_HUMAN_Wan_2019	0.000 [0.000, 1.000]	25
S22A1_HUMAN_Yee_2023_activity	0.333 [0.000, 0.800]	25

0.167

[0.000, 0.483]

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	0.156 [−0.277, 0.562]	25
MK01_HUMAN_Brenan_2016	−0.262 [−0.611, 0.206]	25
SCN5A_HUMAN_Glazer_2019	0.361 [−0.041, 0.646]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	−0.085 [−0.453, 0.292]	25

0.042

[−0.136, 0.241]

100

13 / 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	0.669 [0.429, 0.883]	25
MK01_HUMAN_Brenan_2016	0.449 [0.213, 0.712]	25
SCN5A_HUMAN_Glazer_2019	0.653 [0.397, 0.873]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.507 [0.260, 0.753]	25

0.569

[0.449, 0.687]

100

13 / 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	0.116 [−0.294, 0.650]	25
MK01_HUMAN_Brenan_2016	−0.027 [−0.400, 0.444]	25
SCN5A_HUMAN_Glazer_2019	0.196 [−0.243, 0.603]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	−0.053 [−0.442, 0.368]	25

0.058

[−0.132, 0.306]

100

13 / 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	0.901 [0.720, 0.992]	25
MK01_HUMAN_Brenan_2016	0.409 [0.000, 0.914]	25
SCN5A_HUMAN_Glazer_2019	0.801 [0.376, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.250 [0.039, 0.593]	25

0.590

[0.411, 0.787]

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	0.333 [0.000, 1.000]	25
MK01_HUMAN_Brenan_2016	0.000 [0.000, 0.500]	25
SCN5A_HUMAN_Glazer_2019	0.333 [0.000, 1.000]	25
SRC_HUMAN_Chakraborty_2023_binding-DAS_25uM	0.000 [0.000, 0.000]	25

0.167

[0.000, 0.417]

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	0.640	[0.588, 0.715]	100	15 / 17	0.844	reproduced	unknown	Board
Fmax (biological process, probe) ↑ higher is better · probe	0.499	[0.406, 0.577]	100	3 / 17	0.538	reproduced	unknown	Board
Fmax (cellular component, probe) ↑ higher is better · probe	0.719	[0.667, 0.775]	100	14 / 17	0.756	reproduced	unknown	Board
Micro AUPR (molecular function, probe) ↑ higher is better · probe · secondary	0.623	[0.523, 0.726]	100	—	0.875	reproduced	unknown	Board
Smin (molecular function, probe) ↓ lower is better · bits · probe · secondary	3.584	[2.728, 4.348]	100	—	1.727	reproduced	unknown	Board
Micro AUPR (biological process, probe) ↑ higher is better · probe · secondary	0.465	[0.320, 0.604]	100	—	0.595	reproduced	unknown	Board
Smin (biological process, probe) ↓ lower is better · bits · probe · secondary	6.072	[4.405, 7.407]	100	—	4.842	reproduced	unknown	Board
Micro AUPR (cellular component, probe) ↑ higher is better · probe · secondary	0.712	[0.649, 0.774]	100	—	0.758	reproduced	unknown	Board
Smin (cellular component, probe) ↓ lower is better · bits · probe · secondary	3.445	[2.696, 4.102]	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	0.828	[0.811, 0.847]	1728	13 / 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	0.648	[0.611, 0.686]	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	0.335	[0.151, 0.509]	100	4 / 12	0.410	reproduced	unknown	Board
Spearman ρ (ridge probe) ↑ higher is better · probe	0.032	[−0.167, 0.215]	100	17 / 17	0.297	reproduced	unknown	Board
RMSE, kcal/mol (ridge probe) ↓ lower is better · kcal/mol · probe	1.127	[0.912, 1.330]	100	12 / 17	1.095	reproduced	unknown	Board
Pearson r (ridge probe) ↑ higher is better · probe · secondary	0.060	[−0.111, 0.222]	100	—	0.349	reproduced	unknown	Board
MAE, kcal/mol (ridge probe) ↓ lower is better · kcal/mol · probe · secondary	0.859	[0.714, 1.015]	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	0.656	[0.640, 0.671]	2742	11 / 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	0.475	[0.454, 0.492]	2742	12 / 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	0.580	[0.560, 0.598]	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	0.521 [0.378, 0.654]	100
CASP3_HUMAN_Roychowdhury_2020	0.554 [0.402, 0.677]	100
CCR5_HUMAN_Gill_2023	0.280 [0.079, 0.479]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.596 [0.439, 0.712]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.288 [0.085, 0.466]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.035 [−0.169, 0.222]	100
MLAC_ECOLI_MacRae_2023	0.321 [0.136, 0.485]	100
MTH3_HAEAE_RockahShmuel_2015	0.773 [0.681, 0.838]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.710 [0.564, 0.811]	100
SPG1_STRSG_Olson_2014	0.079 [−0.134, 0.288]	100

0.416

[0.355, 0.464]

1000

5 / 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	0.719 [0.623, 0.816]	100
CASP3_HUMAN_Roychowdhury_2020	0.789 [0.695, 0.870]	100
CCR5_HUMAN_Gill_2023	0.643 [0.531, 0.750]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.805 [0.709, 0.885]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.698 [0.586, 0.797]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.635 [0.502, 0.751]	100
MLAC_ECOLI_MacRae_2023	0.704 [0.603, 0.799]	100
MTH3_HAEAE_RockahShmuel_2015	0.883 [0.809, 0.946]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.904 [0.840, 0.958]	100
SPG1_STRSG_Olson_2014	0.758 [0.599, 0.888]	100

0.754

[0.720, 0.785]

1000

6 / 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	0.300 [0.102, 0.500]	100
CASP3_HUMAN_Roychowdhury_2020	0.501 [0.280, 0.642]	100
CCR5_HUMAN_Gill_2023	0.160 [−0.040, 0.363]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.441 [0.239, 0.621]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.262 [0.046, 0.427]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.223 [0.008, 0.410]	100
MLAC_ECOLI_MacRae_2023	0.240 [0.080, 0.460]	100
MTH3_HAEAE_RockahShmuel_2015	0.515 [0.358, 0.639]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.531 [0.367, 0.708]	100
SPG1_STRSG_Olson_2014	0.231 [0.000, 0.364]	100

0.340

[0.272, 0.389]

1000

7 / 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	0.793 [0.686, 0.925]	100
CASP3_HUMAN_Roychowdhury_2020	0.843 [0.767, 0.917]	100
CCR5_HUMAN_Gill_2023	0.827 [0.613, 0.882]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.812 [0.760, 0.863]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.835 [0.755, 0.917]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.612 [0.446, 0.746]	100
MLAC_ECOLI_MacRae_2023	0.878 [0.814, 0.932]	100
MTH3_HAEAE_RockahShmuel_2015	0.834 [0.759, 0.906]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.877 [0.783, 0.938]	100
SPG1_STRSG_Olson_2014	0.833 [0.704, 0.942]	100

0.814 best

[0.774, 0.842]

1000

1 / 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	0.100 [0.000, 0.400]	100
CASP3_HUMAN_Roychowdhury_2020	0.200 [0.000, 0.455]	100
CCR5_HUMAN_Gill_2023	0.100 [0.000, 0.334]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.000 [0.000, 0.300]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.000 [0.000, 0.250]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.100 [0.000, 0.364]	100
MLAC_ECOLI_MacRae_2023	0.100 [0.000, 0.462]	100
MTH3_HAEAE_RockahShmuel_2015	0.273 [0.000, 0.600]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.400 [0.091, 0.700]	100
SPG1_STRSG_Olson_2014	0.200 [0.000, 0.500]	100

0.147

[0.091, 0.240]

1000

9 / 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	0.401 [0.195, 0.585]	100
CASP3_HUMAN_Roychowdhury_2020	0.604 [0.477, 0.710]	100
CCR5_HUMAN_Gill_2023	0.131 [−0.074, 0.353]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.546 [0.403, 0.662]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.714 [0.571, 0.820]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.501 [0.328, 0.653]	100
MLAC_ECOLI_MacRae_2023	0.382 [0.179, 0.558]	100
MTH3_HAEAE_RockahShmuel_2015	0.535 [0.358, 0.685]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.848 [0.776, 0.899]	100
SPG1_STRSG_Olson_2014	0.782 [0.677, 0.860]	100

0.544

[0.493, 0.585]

1000

11 / 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	0.724 [0.615, 0.819]	100
CASP3_HUMAN_Roychowdhury_2020	0.823 [0.742, 0.895]	100
CCR5_HUMAN_Gill_2023	0.583 [0.472, 0.698]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.801 [0.715, 0.878]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.929 [0.863, 0.981]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.802 [0.683, 0.903]	100
MLAC_ECOLI_MacRae_2023	0.707 [0.611, 0.806]	100
MTH3_HAEAE_RockahShmuel_2015	0.739 [0.602, 0.862]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.949 [0.902, 0.985]	100
SPG1_STRSG_Olson_2014	0.933 [0.881, 0.974]	100

0.799

[0.768, 0.828]

1000

10 / 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	0.340 [0.160, 0.544]	100
CASP3_HUMAN_Roychowdhury_2020	0.421 [0.243, 0.600]	100
CCR5_HUMAN_Gill_2023	0.040 [−0.121, 0.281]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.441 [0.240, 0.590]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.524 [0.328, 0.692]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.401 [0.173, 0.549]	100
MLAC_ECOLI_MacRae_2023	0.320 [0.100, 0.482]	100
MTH3_HAEAE_RockahShmuel_2015	0.328 [0.104, 0.506]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.735 [0.542, 0.817]	100
SPG1_STRSG_Olson_2014	0.403 [0.289, 0.521]	100

0.395

[0.327, 0.446]

1000

10 / 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	0.577 [0.365, 0.857]	100
CASP3_HUMAN_Roychowdhury_2020	0.860 [0.782, 0.923]	100
CCR5_HUMAN_Gill_2023	0.742 [0.568, 0.837]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.813 [0.754, 0.882]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.901 [0.868, 0.967]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.624 [0.485, 0.809]	100
MLAC_ECOLI_MacRae_2023	0.844 [0.749, 0.920]	100
MTH3_HAEAE_RockahShmuel_2015	0.637 [0.390, 0.843]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.942 [0.906, 0.976]	100
SPG1_STRSG_Olson_2014	0.910 [0.855, 0.964]	100

0.785

[0.748, 0.830]

1000

14 / 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	0.200 [0.000, 0.455]	100
CASP3_HUMAN_Roychowdhury_2020	0.200 [0.000, 0.600]	100
CCR5_HUMAN_Gill_2023	0.100 [0.000, 0.385]	100
CP2C9_HUMAN_Amorosi_2021_abundance	0.200 [0.000, 0.500]	100
ILF3_HUMAN_Tsuboyama_2023_2L33	0.200 [0.000, 0.600]	100
KCNE1_HUMAN_Muhammad_2023_expression	0.200 [0.000, 0.500]	100
MLAC_ECOLI_MacRae_2023	0.000 [0.000, 0.400]	100
MTH3_HAEAE_RockahShmuel_2015	0.091 [0.000, 0.455]	100
RPC1_BP434_Tsuboyama_2023_1R69	0.600 [0.200, 0.800]	100
SPG1_STRSG_Olson_2014	0.400 [0.100, 0.700]	100

0.219

[0.163, 0.333]

1000

12 / 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	0.764	[0.747, 0.781]	100	7 / 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	0.610	[0.588, 0.632]	100	7 / 17	0.663	reproduced	overlap Train/test overlap declared by the evaluator: UR100P (UniRef100 + OAS + SCOP)	Board