Summary

Rascore is a tool for analyzing structures of the RAS protein family (KRAS, NRAS, and HRAS). In addition, Rascore can be used to build an updatable database of all available RAS structures in the Protein Data Bank (PDB) with their catalytic switch 1 (SW1) and switch 2 (SW2) loops conformationally classified and their molecular contents annotated (e.g., mutation status, nucleotide state, bound protein, inhibitor site).

Details of our work are provided in the Cancer Research paper, Delineating The RAS Conformational Landscape. We hope that researchers will use Rascore to gain novel insights into RAS biology and drug discovery.

Web Database

A continually updated version of the Rascore database is hosted at http://dunbrack.fccc.edu/rascore/ or https://rascore.streamlitapp.com.

Local Installation

Quickstart commands for environment setup with Anaconda:

conda create -n rascore_env python=3.8

conda activate rascore_env

pip install rascore

conda install -c conda-forge -c schrodinger pymol-bundle=2.5.2

conda install -c conda-forge fpocket=4.0.0

(Additional for macOS: pip install PyQt5)

Usage

Note. Must activate Anaconda enironment with command "conda activate rascore_env" before running.

1. Conformationally classify RAS structures:

rascore -classify [path to coordinate files] -out [output directory path]

• Coordinate files must be provided in mmCIF or PDB formats and numbered according to UniProt scheme. The following inputs are possible:
  • Space-separated list
  • Line-separated list file
  • Tab-separated table file with columns core_path (coordinate path), modelid (optional, model number), chainid (chain identifier), nucleotide_class (optional, nucleotide state)

2. Build the Rascore database from the PDB:

rascore -build [optional, path to pdbaa file] -out [output directory path]

• Takes 1 hour to build from scratch and requires 3 GB of storage.

3. Run the Rascore database application locally:

rascore -gui [optional, path to the Rascore database directory] -out [output directory path]

Authors

Please feel free to contact us with any issues, comments, or questions.

Mitchell Parker

• Email: mip34@drexel.edu or mitchell.parker@fccc.edu
• GitHub: https://github.com/mitch-parker

Roland Dunbrack

• Email: roland.dunbrack@fccc.edu
• GitHub: https://github.com/DunbrackLab

Funding

• NIH NIGMS F30 GM142263 (to M.P.)
• NIH NIGMS R35 GM122517 (to R.D.)

License

Apache License 2.0

Copyright (c) 2022 Mitchell Isaac Parker