This repository contains code corresponding to the paper on bioRxiv titled "Generalized design of sequence-ensemble-function relationships for intrinsically disordered proteins." We provide code corresponding to two example optimizations -- radius of gyration (Rg) and salt sensor optimizations. Additional code is available upon request, and will be made public given sufficient demand.
First, create a fresh conda environment (via mamba or conda). All code was tested with Python version 3.10.13:
mamba create -n <ENV-NAME> python=3.10.13
Next, activate your new environment:
mamba activate ENV-NAME
Then, navigate to this directory and install the required packages. Note that by default, we install the CUDA-compatible version of JAX. If you would like to install the CPU-only version, please remove the corresponding line from requirements.txt and install JAX manually (see here for instructions):
cd path/to/idp-design
pip install -r requirements.txt
Lastly, install sparrow via
pip install git+https://git@github.com/idptools/sparrow.git@a770f78013e6399d992e53921540e559defef94b
All design scripts save results in a specified directory in an output directory. Before you design any IDPs, please cd path/to/idp-design && mkdir output.
To design an IDP with a target Rg, simply run:
python3 -m experiments.design_rg --run-name <RUN-NAME> --seq-length <LENGTH> --target-rg <TARGET-VALUE>
where TARGET-VALUE is the target Rg (in Angstroms) and LENGTH is the length of the IDP. Results will be stored in output/RUN-NAME.
To design an IDP that either expands or contracts depending on changes in salt concentration, run
python3 -m experiments.design_rg_salt_sensor --run-name <RUN-NAME> --seq-length <LENGTH> --salt-lo 150 --salt-hi 450 --mode MODE
where MODE is either expander or contractor and LENGTH is the length of the IDP. Results will be stored in output/RUN-NAME.
By default, the low and high salt concentrations are 150 mM and 450 mM but these can be changed with the --salt-lo and --salt-hi arguments.