AF3Complex yields improved structural predictions of protein complexes.

Area of Science:

• Structural biology
• Computational biology
• Bioinformatics

Background:

• Accurate protein complex structures are crucial for understanding biological pathways.
• Previous work developed AF2Complex, a modification of AlphaFold 2 for protein complexes.
• AlphaFold 3 provides a foundation for advanced structural modeling.

Purpose of the Study:

• Introduce AF3Complex, an enhanced protein complex modeling tool based on AlphaFold 3.
• Evaluate AF3Complex's performance against AlphaFold 3 and other models.
• Assess AF3Complex's capability in modeling challenging protein complex types.

Main Methods:

• Developed AF3Complex by modifying and improving upon AlphaFold 3.
• Benchmarked AF3Complex on a large dataset of protein complexes.
• Evaluated AF3Complex on specific datasets including protein-peptide and antibody-antigen complexes.
• Assessed performance using data from the CASP16 competition.

Main Results:

• AF3Complex demonstrates superior performance compared to AlphaFold 3 on a large benchmark dataset.
• High-fidelity structures were generated for protein-peptide and antibody-antigen complexes.
• AF3Complex predictions would have ranked among the top models in the CASP16 competition.

Conclusions:

• AF3Complex represents a significant advancement in protein complex structure prediction.
• The model shows high accuracy for diverse and challenging protein complex types.
• AF3Complex is a valuable tool for structural biology research and offers competitive performance in structure prediction challenges.