Use AfCycDesign

Protein Design Protein Folding

Overview

AfCycDesign is a cutting-edge deep learning model that leverages the power of AlphaFold2 to facilitate the precise design of macrocyclic peptides. Cyclic peptides have emerged as promising candidates in the realm of therapeutics, but designing them accurately has been a challenge due to limited structural data for molecules of this size. AfCycDesign addresses this gap by modifying the AlphaFold network to predict and design cyclic peptide structures effectively.

Features

  • Utilizes AlphaFold2 for macrocyclic peptide design.
  • Designed for accurate structure prediction and design of cyclic peptides.
  • Addresses challenges in designing cyclic peptides due to limited structural data.
  • Predicts native cyclic peptide structures with high confidence.
  • Impressive accuracy, with 36 out of 49 cases closely matching native structures (pLDDT > 0.85, RMSD < 1.5 Å).
  • Offers computational methods for designing peptide backbones generated by different sampling techniques.
  • Enables de novo design of novel macrocyclic peptides.
  • Extensive structural diversity sampling for cyclic peptides (7-13 amino acids).
  • Identifies around 10,000 unique design candidates with high-confidence folding predictions.
  • Validation through X-ray crystal structures closely matching design models (RMSD < 1.0 Å).
  • Provides a basis for custom-designing peptides for targeted therapeutic applications.
View Paper

Configuration & Options

Model Inputs

The two modes AlphaFold Cyclic Design operates in. De novo will hallucinate a protein to the target specifications and from a template will use that as a reference.

The PDB or AF2 code for your desired template structure. Only enter a value in this option if you are using the template mode. When in template mode, a cyclic protein will be produced using this structure as a reference. Will not work with really big proteins.

The chain you want to use as a template. Only enter a value in this option if you are using the template mode.

A reference amino acid sequence to provide as the initial input to AfCycDesign. This sequence might change during the optimization process.

The length of the cyclic structure you want to generate. Only enter a value in this option if you are using the De novo mode. Will be ignored if Template Sequence is provided.

Ready to submit your job?

Once you're done just hit the submit button below and let us do the rest.

To submit a job please login or register an account.