Prediction of peptide structure: how far are we?

Area of Science:

• Biochemistry
• Computational Biology
• Structural Biology

Background:

• Rational peptide design is hindered by incomplete knowledge of sequence-structure-function relationships.
• Experimental methods like spectroscopy and NMR often yield divergent peptide structure data.
• In silico structure prediction algorithms are crucial for overcoming experimental limitations.

Purpose of the Study:

• To evaluate the utility of the PepLook algorithm for predicting peptide structures and stability.
• To assess PepLook's performance against other algorithms (Pepstr, Robetta) and experimental data.
• To explore how PepLook's indices can inform rational peptide design.

Main Methods:

• Utilized three in silico algorithms: Pepstr, Robetta, and PepLook, to calculate peptide structures from primary sequences.
• Compared calculated structures with experimental data from spectroscopy and Nuclear Magnetic Resonance (NMR).
• Analyzed PepLook's novel indices for evaluating structural polymorphism and stability.

Main Results:

• PepLook and Pepstr accurately predicted structures for peptides with converging experimental data, aligning well with NMR models.
• PepLook's stability index identified potential peptide binding sites.
• PepLook's polymorphism index effectively mapped disordered peptide fragments, explaining discrepancies in divergent experimental data.

Conclusions:

• The PepLook algorithm offers a valuable tool for predicting peptide structure, stability, and polymorphism.
• PepLook's indices provide new avenues for the rational design of peptides by predicting structural behavior.
• Addressing structural polymorphism and assay conditions is key to reconciling computational and experimental peptide structure data.