Monomerizer

Take a PDB structure of a symmetric homo-oligomeric barrel, and monomerize without occluding the pore.

Monomerization of shortened version of PDB 2XQS is shown with only small loops added to monomerize.

Larger modifications can be made aswell such as the large helical linkers for shortened version of PDB 4JPP.

Running Locally as a Single Command

Running requires only a PDB file as input. In the example case we use a shortened version of the 2XQS structure from the RCSB. This file can be found in the repository as tests/pdbs/2XQS_short.pdb.

Make a working directory and move over the input file

1. mkdir docker_tmp
2. cp monomerizer/tests/pdbs/2XQS_short.pdb ./docker_tmp

Make sure the docker deamon is started

1. sudo systemctl start docker

Interactively start the docker image

1. sudo docker run -it --entrypoint /bin/bash -v $(pwd)/tmp_docker/:/workspace --gpus all antiquatedarachnid/monomerizer:latest

Run monomerizer

1. python ../monomerizer/scripts/make_monomer.py 2XQS_short.pdb output --generator-steps 2 --num-designs 1 --linker-lengths 11

Running on the Cloud

In this example we'll run using runpod (https://www.runpod.io/) which will allow easy access to larger and or faster GPUs than you might have locally.

Setup a template for monomerizer

1. Make a new runpod template with the Container Image as antiquatedarachnid/monomerizer:latest
2. Set the Docker Command to /bin/bash

Start the container

1. Launch the container using the runpod WebGUI
2. ssh into the pod using terminal or web-connect options (see RunPod Web GUI buttons)

Copy over your desired input file

1. runpodctl send myfile.pdb on e.g. your local computer and get the code
2. runpodctl receive {code} on the pod

Run monomerizer

1. python ../monomerizer/scripts/make_monomer.py 2XQS_short.pdb output --generator-steps 2 --num-designs 1 --linker-lengths 11

Compress the results and send to your local machine

1. tar -czvf outputs.tar.gz outputs
2. runpodctl send outputs.tar.gz
3. runpotctl receive {code}

Development

If you make changes to the code, you can locally use a the new docker image or publish your own

Building the Container Image

sudo docker build -t monomerizer:latest .

• add --no-cache if needing a full rebuild

Publishing the Docker Image

sudo docker login
sudo docker tag monomerizer:latest your_docker_id/monomerizer:latest
sudo docker push your_docker_id/monomerizer:latest