license: cc-by-4.0
pretty_name: FireProtDB
size_categories:
- 1M<n<10M
task_categories:
- other
language:
- en
tags:
- biology
- proteins
- stability
- mutation
- fireprotdb
- protein-engineering
- train-test-split
- parquet
configs:
- config_name: default
data_files:
- split: train
path:
- data/train-*.parquet
- split: test
path:
- data/test-*.parquet
FireProtDB
Thermostable proteins are crucial in numerous biomedical and biotechnological applications. However, naturally occurring proteins have evolved to function in mild conditions, and laboratory experiments aiming at improving protein stability have proven laborious and expensive. Computational methods overcome this issue by providing a cheap and scalable alternative. Despite significant progress, their reliability is still hindered by the availability of high-quality data. FireProtDB 2.0 (http://loschmidt.chemi.muni.cz/fireprotdb) is a large-scale database aggregating stability data from multiple sources. The second version builds upon its predecessor, retaining its original functionality while introducing a new approach to data storage and maintenance. The new scheme enables the introduction of both absolute and relative data types connected with measurements of wild-types, mutants, protein domains, and de novo designed proteins. Furthermore, while the original database was limited to single-point mutations, more complex data such as insertions, deletions, and multiple-point mutations are now available. As a result, the inclusion of large-scale mutagenesis has increased the size of the database from 16 000 to almost 5 500 000 experiments. Moreover, the updated abstract scheme is fully expandable with any new measurements and annotations without the need for any restructuring. Finally, the tracking of history together with fixed identifiers is in accordance with the FAIR principles.
Dataset Summary
| Metric | Value |
|---|---|
| Rows | 5,465,660 |
| Columns | 99 |
| Source table bytes | 8,466,920,603 |
| Mutant rows | 5,453,252 |
| Sequence rows | 12,408 |
| Substitution events | 5,543,778 |
| Deletion events | 54,296 |
| Insertion events | 50,613 |
Splits
Splits are deterministic by source row id:
sha256("fireprotdb:{row_index}") % 10
Bucket 0 is test; buckets 1 through 9 are train.
| Split | Rows |
|---|---|
train |
4,919,161 |
test |
546,499 |
Source Datasets
| Experiment dataset | Rows |
|---|---|
| Domainome DDG | 4,071,188 |
| MegaScale | 775,235 |
| Domainome FITNESS | 591,671 |
| ProTherm | 27,406 |
| COZYME | 160 |
Common Measurement Columns
The table includes scalar columns for common FireProtDB measurements:
tm, dtm, dg, dg_text, ddg, dh, dcp, dhvh, cm, m_value, trypsin_ml, chymotrypsin_ml, stabilizing, stabilizing_text, domainome_fitness, domainome_fitness_std, domainome_ddg, domainome_ddg_std, reversibility, and state.
If a row has multiple measurements of the same type, the scalar column stores the first value and measurements_json retains the full upstream measurement list.
Loading With datasets
from datasets import load_dataset
ds = load_dataset("LiteFold/FireProtDB")
train = ds["train"]
test = ds["test"]
print(train[0]["protein_name"], train[0]["mutations"], train[0]["tm"])
Load a split directly:
from datasets import load_dataset
train = load_dataset("LiteFold/FireProtDB", split="train")
Filter for rows with DDG measurements:
from datasets import load_dataset
train = load_dataset("LiteFold/FireProtDB", split="train")
ddg_rows = train.filter(lambda row: row["ddg"] is not None)
Stream rows without downloading the full table first:
from datasets import load_dataset
rows = load_dataset("LiteFold/FireProtDB", split="train", streaming=True)
for row in rows:
print(row["row_id"], row["experiment_dataset"], row["measurement_types"])
break
Column Groups
Provenance:
row_id, dataset_id, source_dataset, source_file, source_table, source_sha, row_index, split, subject_type.
Protein and sequence:
entry_id, sequence_id, target_sequence_id, source_sequence_length, target_sequence_length, protein_id, protein_name, organism, isoform, protein_ids, protein_names, organisms, isoforms, uniprot_accessions, interpro_accessions, ec_numbers, megascale_ids, other_references.
Mutation and features:
mutations, substitutions, deletions, insertions, mutation_count, substitution_count, deletion_count, insertion_count, first_position, first_source_aa, first_target_aa, conservation, feature_types, features_json.
Structure:
pdb_ids, afdb_ids, structure_ids, structure_methods, structure_resolution_min, residue_positions, residue_chain_names, residue_secondary_structures, residue_in_pocket_any, residue_in_tunnel_any, residue_asa_mean, residue_bfactor_mean.
Experiment and measurements:
experiment_id, experiment_dataset, ph, measure, method, buffer, buffer_conc, exp_temperature, ion, ion_conc, pdb_chain_mutation, measurement columns, measurement_types, measurement_datasets, annotation_types, annotations_json, measurements_json.
Publication:
publication_id, publication_type, publication_title, publication_year, publication_doi, publication_pmid, publication_journal, publication_url, publication_author_count, publication_authors.
Files
data/train-*.parquet: train split.data/test-*.parquet: test split._MANIFEST.json: source provenance, split counts, and output schema.dataset_summary.json: processing summary and source/measurement counts.scripts/prepare_fireprotdb_dataset.py: script used to generate the flat table.
Source
Derived from LiteFold/FireProtDB, originally sourced from FireProtDB.
License
CC BY 4.0.
Citation
@article{musil2025fireprotdb2,
title = {{FireProtDB} 2.0: large-scale manually curated database of the protein stability data},
author = {Musil, Milos and Borko, Simeon and Planas-Iglesias, Joan and Lacko, David and Rosinska, Monika and Kabourek, Petr and Martins, L{\'i}gia O. and Tataruch, Mateusz and Damborsky, Jiri and Mazurenko, Stanislav and Bednar, David},
journal = {Nucleic Acids Research},
volume = {54},
number = {D1},
pages = {D409--D418},
year = {2026},
publisher = {Oxford University Press},
doi = {10.1093/nar/gkaf1211}
}