SoluProtMut^DB: A manually curated database of protein solubility changes upon mutations

Area of Science:

• Biochemistry
• Structural Biology
• Bioinformatics

Background:

• Protein solubility is crucial for production yields and understanding diseases linked to protein aggregation.
• Limited knowledge exists on protein structural determinants of solubility, with scattered data in literature.
• Existing data lacks comprehensive curation for machine learning applications.

Purpose of the Study:

• To introduce SoluProtMut^DB, the first curated database of protein solubility changes upon mutation.
• To consolidate scattered data and provide a comprehensive resource for protein engineering and disease research.
• To facilitate the development of machine learning models for predicting mutation effects on protein solubility.

Main Methods:

• Compilation of existing literature data on protein solubility and mutations.
• Inclusion of thousands of new data points from recent publications, including deep mutational scanning.
• Manual curation of datasets with corrections and inclusion of experimental conditions affecting solubility.

Main Results:

• The database, SoluProtMut^DB, contains 33,000 measurements for 17,000 protein variants across 103 proteins.
• It integrates previously published datasets with new data, offering extensive coverage.
• The curated data includes various experimental conditions influencing protein solubility.

Conclusions:

• SoluProtMut^DB serves as a vital resource for researchers in protein engineering and computational biology.
• The database enhances understanding of mutation effects on protein solubility and disease connections.
• Improved data quality supports the development of advanced machine learning tools for solubility prediction.