PDBe-SIFTS: an open-source tool for Structure Integration with Function, Taxonomy, and Sequences, featuring improved...

Area of Science:

• Structural biology
• Bioinformatics
• Computational biology

Background:

• Mapping protein sequences to structures is crucial for understanding protein function and mechanisms.
• Existing methods for Structure Integration with Function, Taxonomy and Sequences (SIFTS) rely on internal pipelines.
• Accurate sequence-structure mapping is essential for drug discovery and disease research.

Purpose of the Study:

• To present PDBe-SIFTS, a fully open-source, locally deployable implementation of the SIFTS mapping framework.
• To provide researchers with a tool for fast, scalable, and accurate residue-level sequence-structure mapping.
• To improve structural coherence in alignments and deliver up-to-date, expert-curated quality mappings.

Main Methods:

• Utilizing MMseqs2 for fast and scalable sequence searching.
• Implementing an improved bounded scoring scheme for ranking candidate mappings.
• Employing residue-level mapping refinement based on backbone connectivity.
• Distributing PDBe-SIFTS as a Python package with command-line tools for database building, matching, mapping, and annotation generation.

Main Results:

• MMseqs2 accelerated UniProtKB searches by 22-36 times compared to BLASTP.
• PDBe-SIFTS recovered curated mappings at the top rank in 93.1% of cases.
• The tool demonstrated fast mapping capabilities, improving structural coherence and accuracy.

Conclusions:

• PDBe-SIFTS provides an efficient and accurate solution for residue-level sequence-structure mapping.
• The open-source framework empowers researchers to run, inspect, and extend mappings locally.
• This tool facilitates richer interpretation of biological data and advances in related research fields.