Bridging structure and function: artificial intelligence-based modelling of kidney proteins

Area of Science:

• Structural biology
• Computational biology
• Nephrology

Background:

• Protein structure prediction from amino acid sequences offers insights into function, disease mechanisms, and drug targets.
• Artificial intelligence (AI) algorithms, specifically AlphaFold and RoseTTAFold, have transformed protein modeling.
• AI enables rapid, high-confidence protein structure predictions.

Purpose of the Study:

• To explore the impact of AI-driven protein structure prediction in nephrology.
• To highlight how AI has advanced understanding of renal systems and disease.
• To emphasize the need for integrating AI with experimental validation.

Main Methods:

• Application of AI algorithms (AlphaFold, RoseTTAFold) for protein structure prediction.
• Utilizing predicted structures to analyze renal systems (e.g., podocyte slit diaphragm, membrane transporters, polycystin channels).
• Low-resolution modeling of macromolecular structures for disease mutant analysis and virtual screening.

Main Results:

• Clarified molecular architecture of key renal systems.
• Revealed conformational states of membrane transporters and structural basis of polycystin channelopathies.
• Enabled insights into disease mutant pathogenesis and virtual screening of drugs/toxins.

Conclusions:

• AI models provide significant insights into protein structure and function in nephrology.
• Integration with experimental methods (e.g., cryo-electron tomography) is crucial for capturing dynamic and binding properties.
• Combined AI and experimental approaches will advance understanding of kidney disease pathophysiology and drug discovery.