Surrogate Based Genetic Algorithm Method for Efficient Identification of Low-Energy Peptide Structures.

Area of Science:

• Computational Chemistry
• Molecular Modeling
• Spectroscopy

Background:

• Accurate identification of stable molecular structures is crucial for predicting properties.
• Exploring potential energy surfaces (PES) for molecules is computationally intensive due to the dimensionality bottleneck.
• Efficiently sampling low-lying peptide conformations is essential for interpreting experimental data.

Purpose of the Study:

• To develop and validate an efficient computational approach for sampling realistic low-lying peptide structures.
• To benchmark various computational methods as potential surrogates for a genetic algorithm/density functional theory (GA/DFT) approach.
• To apply the surrogate-guided GA/DFT (sGADFT) method to complex peptide systems and compare results with experimental data.

Main Methods:

• Development of the surrogate-guided genetic algorithm/density functional theory (sGADFT) approach.
• Benchmarking of computational methods including GAFF, AMOEBApro13, PM6, PM7, and DFTB3-D3(BJ) as prescanning surrogates.
• Application of sGADFT to protonated Gly-Pro-Gly-Gly tetrapeptide and doubly protonated gramicidin S.
• Refinement of low-energy structures using density functional theory (DFT) relaxation.

Main Results:

• The sGADFT method successfully identified low-energy peptide minima within hours.
• Refinement of surrogate structures using DFT consistently led to the identification of the most stable conformations.
• Results were validated against high-resolution infrared (IR) spectroscopy data for the tested peptide systems.
• The method demonstrated high efficiency in screening gas-phase peptide structures.

Conclusions:

• The sGADFT method provides a highly efficient computational route for screening realistic low-lying peptide structures.
• This approach is valuable for the interpretation and assignment of experimental IR spectra.
• sGADFT overcomes the dimensionality bottleneck in potential energy surface exploration for peptides.