Peptide docking with AF2 and RosettAfold

This repository holds the raw data and code for reproducing the results and figures of paper
Harnessing protein folding neural networks for peptide-protein docking
by Tomer Tsaban, Julia Varga, Orly Avraham, Ziv Ben-Aharon, Alisa Khramushin, and Ora Schueler-Furman
Nature Communications 13, 176 (2022). https://doi.org/10.1038/s41467-021-27838-9

In the paper, AlphaFold2 was evaluated for peptide-protein docking, on the basis that peptide-protein docking can be seen as a complementing step to protein folding.

The structure of the directory is the following:

|
|_ Data  
      |_ Structures: contains all the models generated with AlphaFold2  
      |_ Source data: all the data used to create the main and supplementary figures in tabular form
      |_ Models: all models generated in this study, sorted by dataset and run types
      |_ Sup tables: supplementary tables (also provided as suppplementary data in the manuscript)

|    
|_ Code  
      |_ Analyses_and_figures: scripts that directly generates plots from the underlying data  
      |_ Running_and_parsing_jobs: all the scripts that prepare the structures for further analysis (removing linker, superimposition, truncation, etc.)  

---

To run a prediction

Use alphafold2_advanced.py script in Code/Running_and_parsing_jobs to run a job. alphafold2_advanced.py -h will give a description with all possible options.

---

If you have any question regarding the data and the code, please send an email to: [email protected]
If you use any of these materials, please cite the paper above.