Automated Pair-wise Peptide-Receptor binding model AnalysIs for Screening Engineered proteins (APPRAISE) is a method that predicts the receptor binding propensity of engineered proteins based on high-precision protein structure prediction tools, such as AlphaFold2-multimer. The APPRAISE Python package includes tools for preparing input files and analyzing the modeled structures.

Author: Xiaozhe Ding (Email: [email protected], [email protected]; Twitter: @DingXiaozhe)

We recommend using APPRAISE remotely by running Colab-APPRAISE notebook on Google Colaboratory, which allows you to access APPRAISE with a web-based interface. This notebook guides users through the APPRAISE process step-by-step, with results stored on Google Drive. No need for a local installation when using this notebook.

The basic service of Google Colaboratory is free, although you can choose paid plans to get more stable access to better hardwares.

How to run Colab-APPRAISE

1. Open Colab-APPRAISE notebook in Google Colaboratory;
2. Go to "File --> save a copy in Drive" to save a copy of your own;
3. Follow the Quick guide on the top of the notebook, and you can start APPRAISing!

Local APPRAISE 1.2 was tested with the following environment:

• MacOS 10.14.6

• Python 3.6.10

• Alphafold-colabfold 2.1.14 (Available here)

• PyMOL 2.3.3 (Schrodinger LLC.)

• Python packages (will be automatically handled by pip):

  • scipy 1.4.1

  • numpy 1.18.2

  • pandas 1.1.5

  • matplotlib 3.2.1

  • seaborn 0.11.2

Installation of APPRAISE locally requires pip. In most cases, pip comes with your Python environment. If not, you can follow the instructions here to install pip.

Install the distribution from PyPI. In the terminal, run:

pip install appraise

Download the repository to your local computer and unzip. In the terminal, change the working folder to the directory containing the appraise package folder and setup.py, and run the following line:

pip install -e .

You can find a few demo notebooks that work locally in the demo folder on GitHub.

Manuscript

Xiaozhe Ding*, Xinhong Chen, Erin E. Sullivan, Timothy F Shay, Viviana Gradinaru*. APPRAISE: Fast, accurate ranking of engineered proteins by receptor binding propensity using structural modeling. Molecular Therapy (2024). * Corresponding authors.

Manuscript-related data

The dataset contains all structural models and sequences used in Ding et al., 2024.

Github repository

The repository contains the latest version of APPRAISE package, Colab-APPRAISE notebook, and demo notebooks.

ColabFold

ColabFold provides a panel of user-friendly tools for structure modeling that are used by APPRAISE.