Antibody characterisation pipeline

Overview

This repository contains code to run the antibody characterisation pipeline described in "Computational design of developable therapeutic antibodies: efficient traversal of binder landscapes and rescue of escape mutations" (see citation below).

Installation

Environment setup

Use the environment.yml file to create a conda environment

conda env create -f environment.yml -n ab-characterisation
conda activate ab-characterisation

Dependencies

Ensure that you have working installs of the following:

1. Rosetta: https://www.rosettacommons.org/demos/latest/tutorials/install_build/install_build

2. ChimeraX: https://www.cgl.ucsf.edu/chimerax/download.html

   a) ensure that you can run the basic ChimeraX script in tests/data:

   ChimeraX --script tests/data/chimera_test_script.py --nogui
   

   b) set environment variable DEBIAN_FRONTEND="noninteractive"

3. ANARCI: https://github.com/oxpig/ANARCI. Note: On MacOS machines, install the hmmer dependency via brew, otherwise via conda.

4. Ensure you have the correct licences for all linked software.

Testing your installation

You can test the installation of the environment using pytest. For this, first set the Rosetta base directory as an environment variable, for example like this:

export ROSETTA_BASE=/path/to/rosetta/rosetta.binary.linux.release-315

Then run pytest

pytest

Which will run an end-to-end example run of the pipeline on a set of 4 antibody sequences (note that depending on your setup this may take 1h).

Running the pipeline

With the conda environment active, the pipeline can be run as follows:

ab-characterisation --input-file tests/data/test_pipeline.csv --rosetta-base-dir $ROSETTA_BASE

(assuming ROSETTA_BASE to have been set as described above).

If you want to multiprocess the pipeline, instead run as

mpiexec -n N_PROCESSES ab-characterisation --input-file tests/data/test_pipeline.csv --rosetta-base-dir $ROSETTA_BASE 

Usage: ab-characterisation [OPTIONS]

╭─ Options ─────────────────────────────────────────────────────────────────────────────────────────────────────────────╮
│ *  --input-file                TEXT     Input .csv file, containing sequence_name, heavy_sequence, light_sequence and │
│                                         reference_complex columns. [required]                                         │
│    --chimera-resolution        FLOAT    Resolution of the map used for alignment within ChimeraX. [default: 6.0]      │
│    --output-dir                TEXT     Directory to which output files are written.                                  │
│                                         [default: ./ab_characterisation_output]                                       │
│    --rosetta-replicates        INTEGER  How many replicates to run for Rosetta characterisation steps. [default: 1]   │
│ *  --rosetta-base-dir          TEXT     Base directory for the Roestta software suite,                                │
│                                         e.g. /path/to/rosetta/rosetta.binary.linux.release-315 [required]             │
│    --top-n                     INTEGER  Top N candidate antibodies to provide from the provided .csv file of          │
│                                         antibodies [default: 10]                                                      │
│    --help                               Show this message and exit.                                                   │
╰────────────────────────────────────────────────

Acknowledgements

The antibody characterisation pipeline was developed by researchers and engineers at Exscientia:

• Frederic Dreyer

• Constantin Schneider

• Aleksandr Kovaltsuk

• Daniel Cutting

• Matthew J. Byrne

• Daniel A. Nissley

• Newton Wahome

• Henry Kenlay

• Claire Marks

• David Errington

• Richard J. Gildea

• David Damerell

• Pedro Tizei

• Wilawan Bunjobpol

• Sachin Surade

• Douglas E. V. Pires

• Charlotte M. Deane

Citation

If you use this code in your research, please cite the following paper:

@article{Computational_design_of_developable_therapeutic_antibodies,
	author = {Dreyer, Fr{\'e}d{\'e}ric A. and Schneider, Constantin and Kovaltsuk, Aleksandr and Cutting, Daniel and Byrne, Matthew J. and Nissley, Daniel A. and Wahome, Newton and Kenlay, Henry and Marks, Claire and Errington, David and Gildea, Richard J. and Damerell, David and Tizei, Pedro and Bunjobpol, Wilawan and Darby, John F. and Drulyte, Ieva and Hurdiss, Daniel L. and Surade, Sachin and Pires, Douglas E. V. and Deane, Charlotte M.},
	title = {Computational design of developable therapeutic antibodies: efficient traversal of binder landscapes and rescue of escape mutations},
	year = {2024},
	doi = {10.1101/2024.10.03.616038},
	eprint = {https://www.biorxiv.org/content/early/2024/10/04/2024.10.03.616038.full.pdf},
	journal = {bioRxiv}
}