Computational modeling of membrane proteins.

Area of Science:

• Structural Biology
• Computational Biology
• Biochemistry

Background:

• Membrane proteins (MPs) are critical drug targets, yet their structures remain underrepresented in the Protein Data Bank (PDB) due to experimental difficulties.
• Less than 2% of PDB entries are MPs, despite over 50% of drugs targeting them, indicating a significant knowledge gap.

Purpose of the Study:

• To review current computational methods for predicting membrane protein structure.
• To highlight recent advancements in predicting both alpha-helical and beta-barrel MP structures.
• To outline future trends in membrane protein structure prediction.

Main Methods:

• Review of computational approaches including secondary structure prediction, transmembrane span and topology prediction.
• Description of comparative modeling, de novo methods, and molecular dynamics (MD) simulations for 3D structure prediction.
• Discussion of how increased experimental structures and correlated mutations improve prediction accuracy.

Main Results:

• Computational methods provide reliable predictions for MP structural features, though challenges remain.
• Recent developments enable more accurate 3D structure predictions for alpha-helical and beta-barrel MPs.
• Availability of more MP structures enhances comparative modeling and knowledge-based scoring functions.

Conclusions:

• Computational prediction is vital for understanding MPs and bridging the gap between their importance and structural knowledge.
• Advances in prediction methods, aided by increased structural data and techniques like MD simulations, are rapidly improving accuracy.
• The field is poised for significant progress in the next decade, driven by these computational strategies.