Protein Stability · LiteFold

MegaScale-Tsuboyama2023

Experimental thermodynamic protein folding stabilities across mutations for training protein stability predictors.

What MegaScale-Tsuboyama2023 is

MegaScale-Tsuboyama2023 provides experimental thermodynamic folding stability measurements for protein variants, compiled from cDNA display proteolysis assays. While structural biology and sequencing reveal sequences and folded coordinates, the thermodynamic forces governing folding are not directly observable from structures alone. This resource captures free energy changes across hundreds of thousands of mutated sequences to aid quantitative study of folding energetics.

The dataset covers variants derived from 331 natural protein domains and 148 de novo designed protein domains ranging from 40 to 72 amino acids in length. It addresses key questions in structural biology, including environmental determinants of amino acid fitness, thermodynamic residue couplings, and the divergence between evolutionary amino acid conservation and physical stability.

Formatted for machine-learning workflows, the collection compiles 299,271 tabular rows into standardised Parquet files. It is structured to train, benchmark, and evaluate computational models that predict free energy changes upon single and multiple amino acid substitutions.

How it was made

Thermodynamic measurements were generated using cDNA display proteolysis, an experimental protocol that measures folding stability across up to 900,000 protein domains in parallel. Proteolytic digestion assays quantify the stability of single amino acid substitutions and selected double mutants across hundreds of small protein folds.

The repository structures these measurements into standardised tabular schemas comprising 269,516 double mutant records alongside site features, natural site statistics, GEMME predictions, and external delta G reference data. Each row provides normalised fields for sequence, mutation position, wild-type and mutant residues, measured delta G values, and complete original records preserved as JSON.

Key facts

Total records: 299,271
Train split records: 269,451
Test split records: 29,820
Primary measured property: Folding free energy (delta G)
Assayed domain length: 40–72 amino acids
Assayed scaffolds: 331 natural and 148 de novo designed protein domains
Dominant record type: Double mutants (269,516 rows)
Data format: Parquet
Modality: Tabular, text

What you can use it for

Predicting mutation stability effects: Training regression algorithms and deep neural networks to predict thermodynamic stability changes (delta G) from primary amino acid sequence and site coordinates.
Investigating epistatic couplings: Analysing non-additive energetic interactions between paired positions in protein domains using the extensive double-mutant stability measurements.
Evaluating computational protein designs: Assessing stability determinants in engineered proteins and benchmarking the success rates of computational de novo protein design methods.
Correlating evolutionary fitness and stability: Comparing observed natural amino acid frequencies and GEMME scores with measured thermodynamic folding stability across conserved protein positions.

Explore the data

The schema, real rows, full-text search and a SQL filter — read in your browser, nothing downloaded.

Explore

Read the real rows without downloading anything

Reading rows…

LiveRead from LiteFold/MegaScale-Tsuboyama2023 at the moment you asked. Nothing is cached or stored — every row above came from that request.

How the columns are distributed

Over 269,451 rows of default/train

a

string_text

min: 3
median: 18
mean: 17.06
max: 19

98.0% empty

aa1

string_label

15%
75%
115%

9.9% empty

aa2

string_label

145%
165%
115%

9.9% empty

aa_seq

string_text

min: 44
median: 72
mean: 68.27
max: 72

99.6% empty

alphatic_contact

string_text

min: 3
median: 18
mean: 17.19
max: 22

94.3% empty

arom_contact

string_text

min: 3
median: 18
mean: 15.83
max: 22

94.3% empty

burial

string_text

min: 3
median: 18
mean: 17.50
max: 22

92.5% empty

d

string_text

min: 3
median: 18
mean: 16.33
max: 19

98.0% empty

dataset_id

string_label

megascale_stability_tsuboyama_2023100%

dboule_mut_name

string_label

1EKL.pdb_dmutv5_4S:54T0%
2KVS.pdb_dmutv5_3Y:31R0%
6EWT.pdb_hnet00%

9.9% empty

ddg_a

string_text

min: 3
median: 19
mean: 17.39
max: 22

98.2% empty

ddg_a_2

string_text

min: 3
median: 19
mean: 17.76
max: 23

94.3% empty

ddg_c

string_text

min: 3
median: 19
mean: 18.86
max: 22

94.3% empty

ddg_d

string_text

min: 3
median: 19
mean: 17.68
max: 22

98.2% empty

ddg_d_2

string_text

min: 3
median: 19
mean: 17.95
max: 22

94.3% empty

ddg_e

string_text

min: 3
median: 19
mean: 17.25
max: 23

98.2% empty

ddg_e_2

string_text

min: 3
median: 19
mean: 17.20
max: 23

94.3% empty

ddg_f

string_text

min: 3
median: 19
mean: 18.21
max: 22

98.2% empty

ddg_f_2

string_text

min: 3
median: 19
mean: 18.42
max: 22

94.3% empty

ddg_g

string_text

min: 3
median: 19
mean: 17.66
max: 22

98.2% empty

ddg_g_2

string_text

min: 3
median: 19
mean: 17.88
max: 22

94.3% empty

ddg_h

string_text

min: 3
median: 19
mean: 18.41
max: 23

98.2% empty

ddg_h_2

string_text

min: 3
median: 19
mean: 18.74
max: 23

94.3% empty

ddg_i

string_text

min: 3
median: 19
mean: 17.86
max: 22

98.2% empty

ddg_i_2

string_text

min: 3
median: 19
mean: 18.05
max: 22

94.3% empty

ddg_k

string_text

min: 3
median: 19
mean: 17.36
max: 23

98.2% empty

ddg_k_2

string_text

min: 3
median: 19
mean: 17.35
max: 23

94.3% empty

ddg_l

string_text

min: 3
median: 19
mean: 17.22
max: 23

98.2% empty

ddg_l_2

string_text

min: 3
median: 19
mean: 17.64
max: 23

94.3% empty

ddg_m

string_text

min: 3
median: 19
mean: 18.44
max: 22

98.2% empty

ddg_m_2

string_text

min: 3
median: 19
mean: 18.71
max: 23

94.3% empty

ddg_n

string_text

min: 3
median: 19
mean: 18.14
max: 23

98.2% empty

ddg_n_2

string_text

min: 3
median: 19
mean: 18.21
max: 23

94.3% empty

ddg_p

string_text

min: 3
median: 19
mean: 17.82
max: 22

98.2% empty

ddg_p_2

string_text

min: 3
median: 19
mean: 18.17
max: 22

94.3% empty

ddg_q

string_text

min: 3
median: 19
mean: 18.05
max: 22

98.2% empty

ddg_q_2

string_text

min: 3
median: 19
mean: 18.32
max: 23

94.3% empty

ddg_r

string_text

min: 3
median: 19
mean: 17.82
max: 23

98.2% empty

ddg_r_2

string_text

min: 3
median: 19
mean: 18.14
max: 23

94.3% empty

ddg_s

string_text

min: 3
median: 19
mean: 17.83
max: 23

98.2% empty

ddg_s_2

string_text

min: 3
median: 19
mean: 18.25
max: 22

94.3% empty

ddg_t

string_text

min: 3
median: 19
mean: 17.91
max: 23

98.2% empty

ddg_t_2

string_text

min: 3
median: 19
mean: 17.93
max: 23

94.3% empty

ddg_v

string_text

min: 3
median: 19
mean: 17.58
max: 22

98.2% empty

ddg_v_2

string_text

min: 3
median: 19
mean: 17.73
max: 23

94.3% empty

ddg_w

string_text

min: 3
median: 19
mean: 18.39
max: 23

98.2% empty

ddg_w_2

string_text

min: 3
median: 19
mean: 18.62
max: 23

94.3% empty

ddg_y

string_text

min: 3
median: 19
mean: 18.15
max: 22

98.2% empty

ddg_y_2

string_text

min: 3
median: 19
mean: 18.34
max: 23

94.3% empty

delta_g

float

min: -6.82
median: 0.90
mean: 0.99
max: 6.13

7.7% empty

deltag

string_text

min: 14
median: 17
mean: 17.30
max: 19

99.6% empty

dg

string_text

min: 3
median: 18
mean: 15.96
max: 23

9.9% empty

dg_hydrophobic_wt_average

string_text

min: 4
median: 18
mean: 18.41
max: 22

98.2% empty

dg_wt_average

string_text

min: 4
median: 18
mean: 18.25
max: 22

98.2% empty

dssp

string_label

92.5% empty

e

string_text

min: 3
median: 18
mean: 16.65
max: 19

98.0% empty

evo_2nd_best

string_text

min: 3
median: 17
mean: 17.07
max: 21

98.2% empty

evo_hydrophobic_normalized

string_text

min: 15
median: 19
mean: 18.67
max: 22

98.2% empty

evo_hydrophobic_wt_average

string_text

min: 4
median: 17
mean: 16.96
max: 20

98.2% empty

evo_normalized

string_text

min: 15
median: 19
mean: 18.64
max: 22

98.2% empty

evo_wt_average

string_text

min: 14
median: 18
mean: 17.51
max: 23

98.2% empty

f

string_text

min: 3
median: 18
mean: 17.31
max: 19

98.0% empty

field

string_text

min: 1
median: 4
mean: 4.35
max: 5

94.3% empty

figure

string_label

Fig490%
Fig36%
Fig52%

func

string_label

98.2% empty

func_2nd_best

string_label

98.2% empty

func_hydrophobic

string_label

98.2% empty

g

string_text

min: 3
median: 18
mean: 16.68
max: 19

98.0% empty

gemme_a

string_text

min: 3
median: 17
mean: 15.60
max: 19

98.2% empty

gemme_d

string_text

min: 3
median: 17
mean: 15.85
max: 19

98.2% empty

gemme_e

string_text

min: 3
median: 17
mean: 15.47
max: 21

98.2% empty

gemme_f

string_text

min: 3
median: 17
mean: 16.15
max: 19

98.2% empty

gemme_g

string_text

min: 3
median: 17
mean: 15.79
max: 19

98.2% empty

gemme_h

string_text

min: 3
median: 17
mean: 16.39
max: 19

98.2% empty

gemme_i

string_text

min: 3
median: 17
mean: 15.89
max: 21

98.2% empty

gemme_k

string_text

min: 3
median: 17
mean: 15.50
max: 19

98.2% empty

gemme_l

string_text

min: 3
median: 17
mean: 15.39
max: 19

98.2% empty

gemme_m

string_text

min: 3
median: 17
mean: 16.35
max: 21

98.2% empty

gemme_n

string_text

min: 3
median: 17
mean: 16.18
max: 19

98.2% empty

gemme_p

string_text

min: 3
median: 17
mean: 16.09
max: 19

98.2% empty

gemme_q

string_text

min: 3
median: 17
mean: 16.11
max: 19

98.2% empty

gemme_r

string_text

min: 3
median: 17
mean: 15.94
max: 19

98.2% empty

gemme_s

string_text

min: 3
median: 17
mean: 15.92
max: 19

98.2% empty

gemme_t

string_text

min: 3
median: 17
mean: 15.99
max: 19

98.2% empty

gemme_v

string_text

min: 3
median: 17
mean: 15.64
max: 19

98.2% empty

gemme_w

string_text

min: 3
median: 17
mean: 16.36
max: 19

98.2% empty

gemme_y

string_text

min: 3
median: 17
mean: 16.08
max: 19

98.2% empty

h

string_text

min: 3
median: 18
mean: 16.94
max: 19

98.0% empty

i

string_text

min: 3
median: 17
mean: 17.25
max: 19

98.0% empty

k

string_text

min: 3
median: 18
mean: 16.70
max: 22

98.0% empty

k50c_a

string_text

min: 3
median: 19
mean: 17.86
max: 23

94.3% empty

k50c_c

string_text

min: 3
median: 19
mean: 18.97
max: 23

94.3% empty

k50c_d

string_text

min: 3
median: 19
mean: 18.08
max: 23

94.3% empty

k50c_e

string_text

min: 3
median: 19
mean: 17.30
max: 22

94.3% empty

k50c_f

string_text

min: 3
median: 19
mean: 18.54
max: 23

94.3% empty

k50c_g

string_text

min: 3
median: 19
mean: 18.03
max: 23

94.3% empty

k50c_h

string_text

min: 3
median: 19
mean: 18.85
max: 23

94.3% empty

k50c_i

string_text

min: 3
median: 19
mean: 18.14
max: 23

94.3% empty

k50c_k

string_text

min: 3
median: 19
mean: 17.45
max: 23

94.3% empty

k50c_l

string_text

min: 3
median: 19
mean: 17.74
max: 23

94.3% empty

k50c_m

string_text

min: 3
median: 19
mean: 18.83
max: 23

94.3% empty

k50c_n

string_text

min: 3
median: 19
mean: 18.33
max: 23

94.3% empty

k50c_p

string_text

min: 3
median: 19
mean: 18.34
max: 22

94.3% empty

k50c_q

string_text

min: 3
median: 19
mean: 18.44
max: 23

94.3% empty

k50c_r

string_text

min: 3
median: 19
mean: 18.27
max: 23

94.3% empty

k50c_s

string_text

min: 3
median: 19
mean: 18.36
max: 23

94.3% empty

k50c_t

string_text

min: 3
median: 19
mean: 18.04
max: 23

94.3% empty

k50c_v

string_text

min: 3
median: 19
mean: 17.82
max: 23

94.3% empty

k50c_w

string_text

min: 3
median: 19
mean: 18.75
max: 23

94.3% empty

k50c_y

string_text

min: 3
median: 19
mean: 18.47
max: 23

94.3% empty

k50t_a

string_text

min: 3
median: 19
mean: 17.82
max: 23

94.3% empty

k50t_c

string_text

min: 3
median: 19
mean: 19.06
max: 23

94.3% empty

k50t_d

string_text

min: 3
median: 19
mean: 18.07
max: 23

94.3% empty

k50t_e

string_text

min: 3
median: 19
mean: 17.29
max: 23

94.3% empty

k50t_f

string_text

min: 3
median: 19
mean: 18.53
max: 23

94.3% empty

k50t_g

string_text

min: 3
median: 19
mean: 18.02
max: 23

94.3% empty

k50t_h

string_text

min: 3
median: 19
mean: 18.82
max: 23

94.3% empty

k50t_i

string_text

min: 3
median: 19
mean: 18.09
max: 23

94.3% empty

k50t_k

string_text

min: 3
median: 19
mean: 17.42
max: 23

94.3% empty

k50t_l

string_text

min: 3
median: 19
mean: 17.68
max: 23

94.3% empty

k50t_m

string_text

min: 3
median: 19
mean: 18.78
max: 23

94.3% empty

k50t_n

string_text

min: 3
median: 19
mean: 18.31
max: 23

94.3% empty

k50t_p

string_text

min: 3
median: 19
mean: 18.35
max: 23

94.3% empty

k50t_q

string_text

min: 3
median: 19
mean: 18.39
max: 23

94.3% empty

k50t_r

string_text

min: 3
median: 19
mean: 18.24
max: 23

94.3% empty

k50t_s

string_text

min: 3
median: 19
mean: 18.34
max: 23

94.3% empty

k50t_t

string_text

min: 3
median: 19
mean: 18.02
max: 23

94.3% empty

k50t_v

string_text

min: 3
median: 19
mean: 17.79
max: 23

94.3% empty

k50t_w

string_text

min: 3
median: 19
mean: 18.75
max: 23

94.3% empty

k50t_y

string_text

min: 3
median: 19
mean: 18.46
max: 23

94.3% empty

l

string_text

min: 3
median: 17
mean: 17.29
max: 19

98.0% empty

m

string_text

min: 3
median: 17
mean: 17.26
max: 19

98.0% empty

meanzero_dg_a

string_text

min: 15
median: 19
mean: 19.01
max: 23

94.3% empty

meanzero_dg_c

string_text

min: 14
median: 18
mean: 18.42
max: 23

94.3% empty

meanzero_dg_d

string_text

min: 15
median: 19
mean: 18.99
max: 23

94.3% empty

meanzero_dg_e

string_text

min: 15
median: 19
mean: 19.03
max: 23

94.3% empty

meanzero_dg_f

string_text

min: 14
median: 19
mean: 18.77
max: 23

94.3% empty

meanzero_dg_g

string_text

min: 14
median: 19
mean: 19.03
max: 23

94.3% empty

meanzero_dg_h

string_text

min: 14
median: 19
mean: 19.19
max: 23

94.3% empty

meanzero_dg_i

string_text

min: 15
median: 19
mean: 18.70
max: 23

94.3% empty

meanzero_dg_k

string_text

min: 15
median: 19
mean: 19.01
max: 23

94.3% empty

meanzero_dg_l

string_text

min: 13
median: 19
mean: 18.68
max: 23

94.3% empty

meanzero_dg_m

string_text

min: 15
median: 19
mean: 18.72
max: 23

94.3% empty

meanzero_dg_n

string_text

min: 15
median: 19
mean: 19.13
max: 23

94.3% empty

meanzero_dg_p

string_text

min: 14
median: 19
mean: 18.85
max: 22

94.3% empty

meanzero_dg_q

string_text

min: 15
median: 19
mean: 19.12
max: 23

94.3% empty

meanzero_dg_r

string_text

min: 15
median: 19
mean: 18.98
max: 23

94.3% empty

meanzero_dg_s

string_text

min: 15
median: 19
mean: 19.15
max: 23

94.3% empty

meanzero_dg_t

string_text

min: 15
median: 19
mean: 19.13
max: 23

94.3% empty

meanzero_dg_v

string_text

min: 14
median: 19
mean: 18.82
max: 23

94.3% empty

meanzero_dg_w

string_text

min: 14
median: 19
mean: 18.78
max: 23

94.3% empty

meanzero_dg_y

string_text

min: 15
median: 19
mean: 18.86
max: 23

94.3% empty

mut_aa

string_label

No distribution reported for this column.

100.0% empty

mutation_name

string_text

min: 6
median: 23
mean: 21.55
max: 37

9.5% empty

n

string_text

min: 3
median: 18
mean: 16.69
max: 22

98.0% empty

name

string_text

min: 6
median: 13
mean: 12.86
max: 31

99.6% empty

negative_contact

string_text

min: 3
median: 18
mean: 17.80
max: 22

94.3% empty

p

string_text

min: 3
median: 18
mean: 15.30
max: 19

98.0% empty

pc1

string_text

min: 15
median: 19
mean: 18.50
max: 22

94.3% empty

pc2

string_text

min: 14
median: 19
mean: 18.60
max: 23

94.3% empty

pc3

string_text

min: 14
median: 19
mean: 18.71
max: 23

94.3% empty

pc4

string_text

min: 14
median: 19
mean: 18.82
max: 23

94.3% empty

pc5

string_text

min: 15
median: 19
mean: 18.91
max: 23

94.3% empty

pdb_id

string_label

1PGA0%
1SHF0%
2M8I0%

99.6% empty

pdb_name

string_label

3DKM.pdb0%
1UFM.pdb0%
2AMI.pdb0%

92.3% empty

polar_contact

string_text

min: 3
median: 18
mean: 17.77
max: 22

94.3% empty

pos

string_label

50%
330%
220%

90.5% empty

position

int

min: 1
median: 29
mean: 29.47
max: 72

90.5% empty

position_1

int

min: 0
median: 9
mean: 9.37
max: 19

9.9% empty

position_2

int

min: 0
median: 9
mean: 9.38
max: 19

9.9% empty

positive_contact

string_text

min: 3
median: 18
mean: 17.91
max: 22

94.3% empty

prev_dg

string_label

0.00%
3.00%
-0.840%

99.6% empty

previous_delta_g

float

min: -0.91
median: 3.59
mean: 3.08
max: 6.10

99.6% empty

protein_id

string_label

HHH4%
3DKM.pdb3%
2MXD.pdb2%

pubmed_id

string_label

313715090%
9819209/12079394/161429140%
195654660%

99.6% empty

q

string_text

min: 3
median: 18
mean: 16.86
max: 19

98.0% empty

r

string_text

min: 3
median: 18
mean: 16.81
max: 19

98.0% empty

raw_row_json

string_text

min: 119
median: 167
mean: 375.23
max: 3,371

raw_wt_aa

string_label

90.5% empty

recon_dg

string_text

min: 3
median: 9
mean: 11.53
max: 23

9.9% empty

reconstructed_delta_g

float

min: -1
median: 0.91
mean: 0.97
max: 5

9.9% empty

record_id

string_text

min: 64
median: 64
mean: 64
max: 64

record_type

string_label

double_mutants90%
site_features6%
natural_sites2%

s

string_text

min: 3
median: 18
mean: 16.88
max: 19

98.0% empty

sasa_sc

string_text

min: 3
median: 17
mean: 15.10
max: 18

92.5% empty

sc_contact

string_text

min: 3
median: 18
mean: 17.55
max: 22

94.3% empty

score_value

float

min: -6.82
median: 0.90
mean: 0.99
max: 6.13

7.7% empty

sequence

string_text

min: 44
median: 72
mean: 68.27
max: 72

99.6% empty

sequence_length

int

min: 44
median: 72
mean: 68.27
max: 72

99.6% empty

source_file

string_label

data/unpacked/labeled/megascale_stability_tsuboyama_2023/megascale_tsuboyama_2023/Data_tables_for_figs/Data_tables_for_figs/dG_for_double_mutants_Fig4.csv90%
data/unpacked/labeled/megascale_stability_tsuboyama_2023/megascale_tsuboyama_2023/Data_tables_for_figs/Data_tables_for_figs/dG_site_feature_Fig3.csv6%
data/unpacked/labeled/megascale_stability_tsuboyama_2023/megascale_tsuboyama_2023/Data_tables_for_figs/Data_tables_for_figs/dG_non_redundant_natural_Fig5.csv2%

source_row_index

int

min: 0
median: 120k
mean: 122k
max: 270k

source_table

string_label

tables/data_unpacked_labeled_megascale_stability_tsuboyama_2023_megascale_tsuboyama_2023_Data_tables_for_figs_Data_tables_for_figs_dG_for_double_mutants_Fig4.csv.jsonl90%
tables/data_unpacked_labeled_megascale_stability_tsuboyama_2023_megascale_tsuboyama_2023_Data_tables_for_figs_Data_tables_for_figs_dG_site_feature_Fig3.csv.jsonl6%
tables/data_unpacked_labeled_megascale_stability_tsuboyama_2023_megascale_tsuboyama_2023_Data_tables_for_figs_Data_tables_for_figs_dG_non_redundant_natural_Fig5.csv.jsonl2%

split_bucket

int

min: 1
median: 5
mean: 5.00
max: 9

t

string_text

min: 3
median: 18
mean: 17.01
max: 19

98.0% empty

table_name

string_label

dG_for_double_mutants_Fig490%
dG_site_feature_Fig36%
dG_non_redundant_natural_Fig52%

target

string_text

min: 3
median: 18
mean: 16.01
max: 23

7.7% empty

thermo_dynamics

string_text

min: 3
median: 19
mean: 17.12
max: 23

9.9% empty

thermodynamic_delta_g

float

min: -5.09
median: 0
mean: 0.01
max: 6

9.9% empty

v

string_text

min: 3
median: 18
mean: 17.21
max: 19

98.0% empty

w

string_text

min: 3
median: 18
mean: 17.24
max: 19

98.0% empty

wt_aa

string_label

90.5% empty

wt_name

string_label

1O6X.pdb0%
2M2J.pdb0%
2MRL.pdb0%

98.2% empty

x_5pc_error

string_text

min: 15
median: 19
mean: 18.63
max: 20

94.3% empty

y

string_text

min: 3
median: 18
mean: 17.20
max: 19

98.0% empty

Strengths and limits

Good at

Contains hundreds of thousands of direct experimental thermodynamic stability values for double-mutant protein sequences.
Normalises variant identifiers, residue positions, sequence strings, and delta G parameters across five distinct source tables.
Preserves raw upstream experimental rows in full JSON columns alongside source table provenance.
Supplies a reproducible deterministic ten-bucket hash split for training and testing.

Watch out for

Experimental assays are restricted to small protein domains between 40 and 72 residues in length.
The published release contains 299,271 rows, representing a selected subset of the total 776,000 curated stabilities reported in the primary study.
Multi-mutation coverage is confined to double mutants, excluding higher-order combinatorial variant sequences.

Questions

Can I look inside MegaScale-Tsuboyama2023 without downloading it?

Yes. This page reads its schema, real rows and per-column statistics in the browser, with full-text search and a SQL filter. Nothing is downloaded to your machine.

What licence is MegaScale-Tsuboyama2023 under?

Its authors declare cc-by-4.0. Read the terms on the original page before reusing the data.

What experimental technique was used to measure stability?

Stabilities were measured via high-throughput cDNA display proteolysis, which assays the protease resistance and thermodynamic folding stability of large protein libraries.

How are the training and test sets partitioned?

Records are assigned using a deterministic SHA-256 hash modulo 10 on the record identifier, placing bucket 0 into the test set and buckets 1 to 9 into the training set.

Which stability metrics are included in the normalised columns?

The schema standardises delta_g, reconstructed_delta_g, thermodynamic_delta_g, previous_delta_g, and generic score_value fields where present in upstream source tables.

What categories of source records are present in this release?

The release aggregates double mutants, site features, natural sites, GEMME natural sites, and external reference delta G measurements.

How to cite it

Tsuboyama, Kotaro, Dauparas, Justas, Chen, Jonathan, Laine, Elodie, Mohseni Behbahani, Yasser, Weinstein, Jonathan J., Mangan, Niall M., Ovchinnikov, Sergey, and Rocklin, Gabriel J. (2023). Mega-scale experimental analysis of protein folding stability in biology and design. Nature, 620(7973), 434–444. doi:10.1038/s41586-023-06328-6

Written from MegaScale-Tsuboyama2023's own dataset card. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.