High-Affinity Peptides for Target Protein Screened in Ultralarge Virtual Libraries

Area of Science:

• Computational chemistry
• Molecular modeling
• Drug discovery

Background:

• High-throughput virtual screening (HTVS) is vital for identifying peptides with high affinity to target proteins for diagnostics and therapeutics.
• Expanding peptide library diversity in HTVS remains a challenge for identifying optimal peptide candidates.

Purpose of the Study:

• To present a de novo design strategy using directed mutation-driven HTVS to evolve large virtual libraries.
• To screen for peptides exhibiting ultrahigh affinities against various protein targets.

Main Methods:

• A computer-generated library of 10^4 random 15-mer peptide scaffolds was created.
• A self-developed algorithm enabled parallelized HTVS using Autodock Vina.
• Top 1% designs underwent 6 generations of random mutation (20% rate), expanding the library theoretically to 10^14 members.

Main Results:

• The strategy was applied to targets including alpha fetoprotein (AFP), SARS-CoV-2 RBD, and norovirus P-domain.
• High-affinity peptides in the nanomolar range were successfully identified for all three targets.
• The approach demonstrated energy savings and high efficiency in peptide design.

Conclusions:

• This directed mutation-driven HTVS strategy efficiently generates vast virtual peptide libraries.
• It enables the cost-effective development of effective high-affinity peptides for health and environmental applications.
• The method offers new opportunities for peptide-based diagnostics and therapeutics.