# Start with a copy of the cuda image maintained by Nvidia to avoid FROM nvcr.io/nvidia/cuda:11.4.2-base-ubuntu20.04 # Install basic tools RUN apt-get update && apt-get install -y \ wget \ curl \ unzip # Install miniconda and awscli RUN curl -L -o ~/miniconda.sh https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh \ && chmod +x ~/miniconda.sh \ && ~/miniconda.sh -b -p /opt/conda \ && rm ~/miniconda.sh \ && /opt/conda/bin/conda update conda \ && /opt/conda/bin/conda install -c conda-forge awscli # Download and unzip v1.1 of the RoseTTAFold repository, available at # https://github.com/RosettaCommons/RoseTTAFold RUN wget https://github.com/RosettaCommons/RoseTTAFold/archive/refs/tags/v1.1.0.zip \ && unzip v1.1.0.zip \ && mv RoseTTAFold-1.1.0 /RoseTTAFold \ && rm v1.1.0.zip WORKDIR /RoseTTAFold # Install lddt, cs-blast, and libgomp1 RUN ./install_dependencies.sh RUN /opt/conda/bin/conda env create -f RoseTTAFold-linux.yml \ && /opt/conda/bin/conda clean -ya RUN apt-get install libgomp1 # Add the AWS-RoseTTAFold scripts COPY run_aws_data_prep_ver.sh . COPY run_aws_predict_ver.sh . COPY download_ref_data.sh . # Clean up unecessary files to save space RUN rm -rf \ example \ folding \ *.gz \ *.zip \ *.yml \ install_dependencies.sh # Create a directory to mount the FSx Lustre file system with ref data VOLUME /fsx # Activate conda\ RUN ["/bin/bash", "-c", \ "/opt/conda/bin/activate", \ "/opt/conda/bin/conda init bash", \ "source $HOME/.bashrc"] ENV PATH /opt/conda/bin:$PATH # Define the default run command. Batch will overwrite this at run time. CMD ["/bin/bash"]