De novo design of proteins using Rosetta and deep learning methods (workshop report)
With the help of Instruct.SI, Dr. Ajasja Ljubetič from National Institute of Chemistry organised a one-day workshop on protein design entitled “De novo design of proteins using Rosetta and deep learning methods” that took place on March 13th, 2024. More than 80 researchers applied and 50 were accepted. The topics covered ranged from basic PyRosetta knowledge to state-of-the-art deep learning methods such as AlphaFold2, ProteinMPNN and RoseTTAFold diffusion.
The workshop was very interactive with Jupyter notebooks that participants could run on Google Colab, and our four excellent teaching assistants (Eva Rajh, Tadej Satler, Alina Konstantinova and Dr. Federico Alberto Olivieri) made sure everything ran smoothly.
48% of participants were very satisfied with the workshop and 31% found it very useful for their work. Some quotes from participants:
- My favorite part was RF diffusion, I think it makes creating your own proteins based on your specification really user-friendly, so I think it’s great even for researchers whose main focus isn’t de novo protein design.
- I enjoyed the whole workshop. It was really informative and exciting.
- When I managed to write the program that created a coiled coil. It was a good feeling to immediately see a visual result.
Photo credits: Dr. Andrej Perdih