AlphaSync is an enhanced AlphaFold structure database synchronized with UniProt

Area of Science:

• Structural biology
• Bioinformatics
• Computational biology

Background:

• Accurate protein structure prediction is crucial for biological understanding and biomedical research.
• Keeping protein structure models synchronized with evolving sequence databases is a significant challenge.

Purpose of the Study:

• To introduce AlphaSync, a resource complementing the AlphaFold Protein Structure Database.
• To provide up-to-date, UniProt-synchronized protein structural models and annotations.

Main Methods:

• Developed AlphaSync (alphasync.stjude.org) to synchronize protein structure models with UniProt.
• Generated 2.6 million structural models for updated proteins and isoforms across 925 species.
• Achieved complete proteome coverage for 42 key species.

Main Results:

• AlphaSync offers 2.6 million UniProt-synchronized structural models, including predictions for 40,016 updated proteins.
• Provides complete, up-to-date proteome coverage for 42 species, including humans and pathogens.
• Includes residue-level annotations (solvent accessibility, disorder) and over 4.7 billion atom-level noncovalent contacts.

Conclusions:

• AlphaSync facilitates protein structure-function relationship studies and variant assessment.
• Enables large-scale and detailed protein research through up-to-date models and annotations.
• Supports machine learning tasks such as protein design with comprehensive data.