diffdock

utils/docking/diffdock/diffdock-tool/.bumpversion.cfg

@@ -0,0 +1,29 @@
+[bumpversion]
+current_version = 0.1.0
+commit = False
+tag = False
+parse = (?P<major>\d+)\.(?P<minor>\d+)\.(?P<patch>\d+)(\-(?P<release>[a-z]+)(?P<dev>\d+))?
+serialize = 
+	{major}.{minor}.{patch}-{release}{dev}
+	{major}.{minor}.{patch}
+
+[bumpversion:part:release]
+optional_value = _
+first_value = dev
+values = 
+	dev
+	_
+
+[bumpversion:part:dev]
+
+[bumpversion:file:pyproject.toml]
+search = version = "{current_version}"
+replace = version = "{new_version}"
+
+[bumpversion:file:VERSION]
+
+[bumpversion:file:README.md]
+
+[bumpversion:file:plugin.json]
+
+[bumpversion:file:src/polus/mm/utils/diffdock/__init__.py]

utils/docking/diffdock/diffdock-tool/.dockerignore

@@ -0,0 +1,4 @@
+.venv
+out
+tests
+__pycache__

utils/docking/diffdock/diffdock-tool/.gitattributes

@@ -0,0 +1,5 @@
+*.pdb filter=lfs diff=lfs merge=lfs -text
+*.pdbqt filter=lfs diff=lfs merge=lfs -text
+*.mol2 filter=lfs diff=lfs merge=lfs -text
+*.xlsx filter=lfs diff=lfs merge=lfs -text
+*.sdf filter=lfs diff=lfs merge=lfs -text

utils/docking/diffdock/diffdock-tool/.gitignore

@@ -0,0 +1 @@
+poetry.lock

utils/docking/diffdock/diffdock-tool/CHANGELOG.md

@@ -0,0 +1,5 @@
+# CHANGELOG
+
+## 0.1.0
+
+Initial release.

utils/docking/diffdock/diffdock-tool/Dockerfile_diffdock_gpu

@@ -0,0 +1,78 @@
+# docker build -f Dockerfile_diffdock_gpu -t mrbrandonwalker/diffdock-gpu-tool .
+# Stage 1: Build Environment Setup
+FROM nvidia/cuda:11.7.1-devel-ubuntu22.04 as builder
+
+RUN apt-get update -y && apt-get install -y wget curl git tar bzip2 unzip && rm -rf /var/lib/apt/lists/*
+
+# Create a user
+ENV APPUSER="appuser"
+ENV HOME=/home/$APPUSER
+RUN useradd -m -u 1000 $APPUSER
+USER $APPUSER
+WORKDIR $HOME
+
+ENV ENV_NAME="diffdock"
+ENV DIR_NAME="DiffDock"
+
+# Install micromamba
+RUN curl -Ls https://micro.mamba.pm/api/micromamba/linux-64/latest | tar -xj bin/micromamba
+ENV PATH=$HOME/bin:$HOME/.local/bin:$PATH
+
+# Copy and create Conda environment
+ENV ENV_FILE_NAME=environment.yml
+RUN git clone https://github.com/gcorso/DiffDock.git $DIR_NAME
+WORKDIR $HOME/$DIR_NAME
+RUN ~/bin/micromamba env create --file $ENV_FILE_NAME && ~/bin/micromamba clean -afy --quiet
+
+# Download models
+# These should download automatically on first inference
+# RUN curl -L -o diffdock_models_v1.1.zip "https://www.dropbox.com/scl/fi/drg90rst8uhd2633tyou0/diffdock_models.zip?rlkey=afzq4kuqor2jb8adah41ro2lz&dl=1" \
+#     && mkdir -p $HOME/$DIR_NAME/workdir \
+#     && unzip diffdock_models_v1.1.zip -d $HOME/$DIR_NAME/workdir
+
+
+# Stage 2: Runtime Environment
+FROM nvidia/cuda:11.7.1-runtime-ubuntu22.04
+
+# Create user and setup environment
+ENV APPUSER="appuser"
+ENV HOME=/home/$APPUSER
+RUN useradd -m -u 1000 $APPUSER
+USER $APPUSER
+WORKDIR $HOME
+
+ENV ENV_NAME="diffdock"
+ENV DIR_NAME="DiffDock"
+
+# Copy the Conda environment and application code from the builder stage
+COPY --from=builder --chown=$APPUSER:$APPUSER $HOME/micromamba $HOME/micromamba
+COPY --from=builder --chown=$APPUSER:$APPUSER $HOME/bin $HOME/bin
+COPY --from=builder --chown=$APPUSER:$APPUSER $HOME/$DIR_NAME $HOME/$DIR_NAME
+COPY --from=builder --chown=$APPUSER:$APPUSER $HOME/$DIR_NAME $HOME/$DIR_NAME
+WORKDIR $HOME/$DIR_NAME
+
+# Set the environment variables
+ENV MAMBA_ROOT_PREFIX=$HOME/micromamba
+ENV PATH=$HOME/bin:$HOME/.local/bin:$PATH
+RUN micromamba shell init -s bash --root-prefix $MAMBA_ROOT_PREFIX
+
+# Precompute series for SO(2) and SO(3) groups
+RUN micromamba run -n ${ENV_NAME} python utils/precompute_series.py
+
+# Expose ports for streamlit and gradio
+EXPOSE 7860 8501
+
+# shell file to copy cached files, run diffdock and remove large cached files after execution
+COPY diffdock_cmds.sh $HOME/$DIR_NAME
+
+# Switch to root user to change ownership and permissions
+USER root
+
+# Change ownership to root and set permissions
+RUN chown appuser:appuser $HOME/$DIR_NAME/diffdock_cmds.sh
+RUN chmod 777 $HOME/$DIR_NAME/diffdock_cmds.sh
+USER appuser
+
+# Default command
+CMD ["sh", "-c", "micromamba run -n ${ENV_NAME} python utils/print_device.py"]
+ENTRYPOINT []

utils/docking/diffdock/diffdock-tool/README.md

@@ -0,0 +1,22 @@
+# diffdock (0.1.0)
+
+DiffDock Diffusion based protein ligand docking
+
+## Options
+
+This plugin takes     11     input arguments and 4 output argument:
+
+| Name          | Description             | I/O    | Type   | Default |
+|---------------|-------------------------|--------|--------|---------|
+| protein_path | Protein input file | Input | File | File |
+| ligand_path | Ligand input file | Input | File | File |
+| inference_steps | Number of inference steps for diffusion | Input | int | int |
+| samples_per_complex | Number of pose samples to generate per complex | Input | int | int |
+| batch_size | Batch size | Input | int | int |
+| out_dir | Output directory to save poses | Input | string | string |
+| model_dir | Input model directory to use | Input | string | string |
+| confidence_model_dir | Input confidence model directory | Input | string | string |
+| complex_name | Name of complex | Input | string | string |
+| max_confident_pose | Highest confident pose | Output | File | File |
+| output_files | The output poses | Output | File[] | File[] |
+| execution_time | Time to run DiffDock | Output | float | float |

utils/docking/diffdock/diffdock-tool/VERSION

@@ -0,0 +1 @@
+0.1.0

utils/docking/diffdock/diffdock-tool/diffdock_0@1@0.cwl

@@ -0,0 +1,140 @@
+#!/usr/bin/env cwl-runner
+cwlVersion: v1.0
+
+class: CommandLineTool
+
+label: DiffDock Diffusion based protein ligand docking
+
+doc: |-
+  DiffDock Diffusion based protein ligand docking
+
+baseCommand: ["bash", "/home/appuser/DiffDock/diffdock_cmds.sh"]
+
+
+hints:
+  cwltool:CUDARequirement:
+    cudaVersionMin: "11.7"
+    cudaComputeCapability: "3.0"
+    cudaDeviceCountMin: 1
+    cudaDeviceCountMax: 1
+  DockerRequirement:
+    dockerPull: mrbrandonwalker/diffdock-gpu-tool@sha256:ae2a6de920e8d250cabcf86a20f9e7be28503c0d70483635a7639392bed7f4c2
+
+requirements:
+  InlineJavascriptRequirement: {}
+
+inputs:
+
+  protein_path:
+    type: File
+    format: edam:format_1476
+    inputBinding:
+      prefix: --protein_path
+
+  ligand_path:
+    type: File
+    format: edam:format_3814
+    inputBinding:
+      prefix: --ligand_description
+
+  inference_steps:
+    label: number of reverse diffusion steps
+    type: int?
+    inputBinding:
+      prefix: --inference_steps
+    default: 20
+
+  samples_per_complex:
+    label: Number of sample poses to generate per complex
+    type: int?
+    inputBinding:
+      prefix: --samples_per_complex
+    default: 40
+
+  batch_size:
+    label: input batch size for neural net
+    type: int?
+    inputBinding:
+      prefix: --batch_size
+    default: 1
+
+  out_dir:
+    label: where output from diffdock is saved
+    type: string?
+    inputBinding:
+      prefix: --out_dir
+    default: results/
+
+  model_dir:
+    label: directory of DiffDock score model from paper
+    type: string?
+    inputBinding:
+      prefix: --model_dir
+    default:  /DiffDock/workdir/paper_score_model/
+
+  confidence_model_dir:
+    label: directory of DiffDock confidence model from paper
+    type: string?
+    inputBinding:
+      prefix: --confidence_model_dir
+    default:  /DiffDock/workdir/paper_confidence_model
+
+  complex_name:
+    label: name of folder with pose outputs that will be saved under out_dir folder
+    type: string?
+    inputBinding:
+      prefix: --complex_name
+    default:  outputs
+
+  output_files:
+    type: string?
+
+  max_confident_pose:
+    type: string?
+
+outputs:
+
+  max_confident_pose:
+    type: File
+    outputBinding:
+      # the diffdock developers copy only the top ranked pose to a new file rank1.sdf
+      glob: $(inputs.out_dir)/$(inputs.complex_name)/rank1.sdf
+    format: edam:format_3814
+
+  output_files:
+    type: File[]
+    outputBinding:
+      # all other output files besides rank1.sdf have confidence information in them rank*_confidence*.sdf
+      glob: $(inputs.out_dir)/$(inputs.complex_name)/rank*_confidence*.sdf
+    format: edam:format_3814
+
+  stderr:
+    type: File
+    outputBinding:
+      glob: stderr
+
+  execution_time:
+    label: Time to run DiffDock
+    doc: |-
+      Time to run DiffDock
+    type: float
+    outputBinding:
+      glob: stderr
+      loadContents: true
+      outputEval: |
+        ${
+          // the time command outputs to stderr and not to stdout
+          // example output below, parse the float value of seconds (first item in line)
+          // 1it [00:41, 41.03s/it]
+          // 78.909
+          return self[0].contents.split("\n").map(str => parseFloat(str)).reverse().find(num => !isNaN(num));
+        }
+
+stderr: stderr
+
+$namespaces:
+  edam: https://edamontology.org/
+  cwltool: http://commonwl.org/cwltool#
+
+$schemas:
+- https://raw.githubusercontent.com/edamontology/edamontology/master/EDAM_dev.owl

utils/docking/diffdock/diffdock-tool/diffdock_cmds.sh

@@ -0,0 +1,7 @@
+#!/bin/bash -e
+# copy cached files to current directory
+cp /DiffDock/.*.npy .
+TIMEFORMAT=%R && time python /DiffDock/inference.py "$@"
+# need to remove large files otherwise cachedir folder will be 3GB each!!
+rm .*.npy
+rm -r .cache/