Associating protein residues in the literature with structural data

Area of Science:

• Structural biology
• Bioinformatics
• Computational biology

Background:

• Protein structures are vital for understanding cellular functions and disease mechanisms.
• Accessing and validating protein structure data linked to scientific literature is challenging for researchers.
• Current methods require multiple software packages, hindering efficient data exploration.

Purpose of the Study:

• To develop an integrated software tool for associating protein residue mentions in text with their corresponding 3D structures.
• To provide researchers with a single-view platform for exploring protein structure, publication context, and experimental evidence.
• To enhance the accessibility and validation of protein structure-function relationships.

Main Methods:

• Implementation of an artificial intelligence (AI) and text-mining approach.
• Development of algorithms for annotation extraction and downstream processing.
• Integration with databases like the International Union of Crystallography (IUCr) and the Protein Data Bank in Europe (PDBe) for dissemination and visualization.

Main Results:

• A novel software tool successfully links textual mentions of protein residues to their specific locations in 3D protein structures.
• The tool enables a unified view of a protein residue within its publication context and experimental data.
• Demonstrated model implementation, annotation extraction, and visualization processes.

Conclusions:

• The developed tool offers a significant advancement in accessing and interpreting protein structure-function data.
• It streamlines the process for researchers, especially non-experts, to assess the validity of conclusions based on experimental results.
• Future applications include aiding journal reviewers and authors in validating annotations and improving data curation.