Proteins · Database
FireProtDB
A curated database of protein stability measurements for single-point mutants to train and test stability models.
What FireProtDB is
FireProtDB is a curated repository of protein thermostability data for single-point mutants, developed by Loschmidt Laboratories. It addresses the lack of experimental stability data that limits the training, testing, and predictive accuracy of computational protein design tools.
Naturally occurring proteins generally evolve to function under mild physiological conditions and frequently fail to withstand harsh industrial or biotechnological environments. By consolidating stability measurements across multiple experimental efforts, the resource supports research into mutational effects and the engineering of stabilised proteins for biotechnology and biomedicine.
How it is organised
The database integrates published datasets from ProTherm, Megascale, and Human domainome, alongside entries extracted directly from scientific literature. It also incorporates laboratory measurements produced by Loschmidt Laboratories and external collaborators within the COZYME initiative. Additional annotations are gathered from VariBench and HotSpot Wizard.
Data can be retrieved and queried through a graphical user interface. The platform supports interactive viewing of individual proteins and mutational records, as well as advanced search and filtering functions designed to export customised datasets for machine learning workflows.
Key facts
- Records
- 24k measurements
- Data scope
- Thermostability measurements for single-point mutants
- Integrated data sources
- ProTherm, Megascale, Human domainome, literature, and COZYME initiative
- Annotation sources
- VariBench and HotSpot Wizard
What you can use it for
- Machine learning dataset generation
- Constructing custom, filtered datasets of single-point mutational data for training and evaluating protein stability prediction models.
- Single-point mutation lookup
- Interactively exploring experimental thermostability measurements for specific amino acid substitutions and protein targets.
- Protein engineering
- Analysing the stability impacts of point mutations to guide the design of proteins that tolerate harsh operational environments.
Strengths and limits
Good at
- Gathers manually curated stability measurements from disparate public datasets, literature, and internal laboratory experiments into a single resource.
- Enriches stability measurements with annotations derived from VariBench and HotSpot Wizard.
- Provides dedicated filtering and search features tailored for creating machine learning-friendly datasets.
Watch out for
- Coverage is restricted to single-point mutations and does not document multiple-point mutations or complex structural variants.
Questions
Can I search FireProtDB without writing code?
FireProtDB is catalogued here but not yet queried live, so the agent can point you to it but cannot search it for you yet.
Can I reuse data from FireProtDB?
FireProtDB publishes its own terms of use; read them on https://loschmidt.chemi.muni.cz/fireprotdb before reusing its records.
What kind of data does FireProtDB store?
FireProtDB stores manually curated protein thermostability measurements for single-point mutants.
Where do the measurements come from?
The records combine datasets from ProTherm, Megascale, and Human domainome, literature extractions, and experimental measurements from Loschmidt Laboratories and the COZYME initiative.
Can FireProtDB data be exported for computational modelling?
Yes, the interface provides advanced filtering and search features specifically designed to produce machine learning-friendly datasets.
How to cite it
Musil, M., Borko, S., Planas-Iglesias, J., Lacko, D., Rosinska, M., Kabourek, P., Martins, L.O., Tataruch, M., Damborsky, J., Mazurenko, S., Bednar, D., 2025 : FireProt DB 2.0: large-scale manually curated database of the protein stability data. Nucleic Acids Research .
Written from FireProtDB's own website and reference material. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.