- We will outline basic methods of working with protein design using pyRosetta, ProteinMPNN and RFDiffusion.
- Duration: 90 minutes
- Objectives: be able to
- Understand how to setup modeling with Rosetta
- Understand how to setup protein folding simulations in Rosetta
- Understand how to vary a binding sequence using ProteinMPNN
- Understand how to generate new protein scaffolds using RFDiffusion
- PyRosetta Workshops.
- The github repo (PyRosetta) is here
- ProteinMPNN basic and example
- RFDiffusion
- Easiest way is to use Google Colab Google Drive - see tutorial for initialization instructions. However, this will require a lengthy download.
- Or you need local access to a Jupyter notebook evironment with pyRosetta installed, see here
- Or use SandBox cluster rosetta conda environment.
Take a structure of interest (with 2 or more protein chains):
- Using pyRosetta, analyze energy between any two close residues of separate chains, output various energy terms.See 03.02-Analyzing-energy-between-residues.ipynb
- (со звуздочкой) Using pyRosetta, take an arbitrary sequence of 10 amino acids. Try to do folding. Explore different lowest conformations depeding on the number of Monte-Carlo steps. Visualize results. See 04.01-Basic-Folding-Algorithm.ipynb
- Use RFDiffusion to generate a new binder scaffold for the target protein. Use ProteinMPNN to generate the amino acid residues of the new binder. Use AlphaFold to check whether a complex similar to the original structure is formed.
- Consult with the seminar protocol/recording
- Ask questions in Slack