Configure bioemu for WES deployment

- Update main.nf: Harbor container image, PVC mount paths, remove containerOptions and stageInMode - Update nextflow.config: Add k8s/k8s_gpu profiles, minimal config for WES injection - Update params.json: Correct paths for eureka PVC - Update Dockerfile.wes: CUDA base image for GPU support
2026-03-26 14:11:57 +01:00
parent ca4ceae21e
commit f776745722
4 changed files with 120 additions and 107 deletions
--- a/Dockerfile.wes
+++ b/Dockerfile.wes
@@ -1,15 +1,14 @@
-FROM python:3.11-slim
+FROM nvidia/cuda:12.2.0-runtime-ubuntu22.04
+
+ENV DEBIAN_FRONTEND=noninteractive
+
+RUN apt-get update && apt-get install -y \
+    wget git python3 python3-pip build-essential curl \
+    && rm -rf /var/lib/apt/lists/*

 WORKDIR /opt/bioemu

-# Install minimal system deps
-RUN apt-get update && apt-get install -y --no-install-recommends \
-    gcc g++ && rm -rf /var/lib/apt/lists/*
-
-# Install bioemu with CPU-only PyTorch first (smaller)
-# The k8s GPU nodes will have CUDA drivers available
-RUN pip install --no-cache-dir --upgrade pip && \
-    pip install --no-cache-dir torch --index-url https://download.pytorch.org/whl/cpu && \
+RUN python3 -m pip install --upgrade pip && \
    pip install --no-cache-dir bioemu mdtraj scikit-learn pandas matplotlib seaborn

 RUN mkdir -p /opt/bioemu/scripts/ /data /results && chmod -R 777 /data /results
--- a/main.nf
+++ b/main.nf
@@ -1,23 +1,22 @@
 #!/usr/bin/env nextflow
 nextflow.enable.dsl=2

-// Define parameters
-params.complex_name = "hgh_mab1"  // Default complex name
-params.protein1_fasta = "/mnt/OmicNAS/private/old/olamide/bioemu/input/complexes/hgh.fasta"
-params.protein2_fasta = "/mnt/OmicNAS/private/old/olamide/bioemu/input/complexes/mab1.fasta"
-params.exp_dG = -13.1  // kcal/mol from experimental database
-params.outdir = "/mnt/OmicNAS/private/old/olamide/bioemu/output"
-params.cache_dir = "/mnt/OmicNAS/private/old/olamide/bioemu/cache"
+// Define parameters — PVC mount paths for k8s execution
+params.protein1_fasta = '/omic/eureka/bioemu/input/trp_cage.fasta'
+params.protein2_fasta = '/omic/eureka/bioemu/input/villin_headpiece.fasta'
+params.complex_name = "protein_complex"
+params.exp_dG = -10.0  // kcal/mol (placeholder experimental value)
+params.outdir = '/omic/eureka/bioemu/output'
+params.cache_dir = '/tmp/bioemu_cache'

 // Parameters for structure generation and analysis
-params.num_samples = 20
+params.num_samples = 10
 params.batch_size = 5
 params.temperature = 300
 params.n_clusters = 5

 process GENERATE_STRUCTURE {
-    container 'bioemu:latest'
-    containerOptions '--rm --gpus all -v /mnt:/mnt'
+    container 'harbor.cluster.omic.ai/omic/bioemu:latest'
    publishDir "${params.outdir}/${params.complex_name}/${protein_id}", mode: 'copy'

    input:
@@ -31,6 +30,9 @@ process GENERATE_STRUCTURE {
    # Extract sequence from FASTA
    SEQUENCE=\$(grep -v ">" ${fasta} | tr -d '\\n')

+    # Create cache dir
+    mkdir -p ${params.cache_dir}
+
    # Run BioEmu
    python3 -m bioemu.sample \\
        --sequence "\${SEQUENCE}" \\
@@ -46,8 +48,7 @@ process GENERATE_STRUCTURE {
 }

 process CALCULATE_BINDING {
-    container 'bioemu:latest'
-    containerOptions '--rm --gpus all -v /mnt:/mnt -v /data:/data'
+    container 'harbor.cluster.omic.ai/omic/bioemu:latest'
    publishDir "${params.outdir}/${params.complex_name}/analysis", mode: 'copy'

    input:
@@ -63,8 +64,8 @@ process CALCULATE_BINDING {

    script:
    """
-    # Run binding energy calculation with the existing script
-    python3 /data/olamide/fresh-bioemu2/scripts/calculate_binding.py \\
+    # Run binding energy calculation
+    python3 /opt/bioemu/scripts/calculate_binding.py \\
        --protein1_topology ${protein1_topology} \\
        --protein1_samples ${protein1_samples} \\
        --protein2_topology ${protein2_topology} \\
@@ -119,7 +120,7 @@ workflow {
                      .map { it -> tuple(it[1], it[2]) }
                      .first()

-    // Calculate binding energy (direct script reference)
+    // Calculate binding energy
    CALCULATE_BINDING(
        protein1_files[0],  // protein1_topology.pdb
        protein1_files[1],  // protein1_samples.xtc
--- a/nextflow.config
+++ b/nextflow.config
@@ -1,38 +1,55 @@
-// Manifest for Nextflow metadata
 manifest {
-    name = 'BioEmu-Nextflow'
-    author = 'Generated from BioEmu repository'
-    homePage = 'https://github.com/microsoft/bioemu'
-    description = 'Nextflow pipeline for BioEmu - Biomolecular Emulator for protein structure sampling'
+    name = 'bioemu'
+    author = 'Olamide'
+    description = 'BioEmu - Biomolecular Emulator for protein structure sampling and binding energy analysis'
    mainScript = 'main.nf'
    version = '1.0.0'
 }

-// Global default parameters
 params {
-    fasta = "/mnt/OmicNAS/private/old/olamide/bioemu/input/villin_headpiece.fasta"
-    outdir = "/mnt/OmicNAS/private/old/olamide/bioemu/output"
-    cache_dir = "/mnt/OmicNAS/private/old/olamide/bioemu/cache"
+    protein1_fasta = '/omic/eureka/bioemu/input/trp_cage.fasta'
+    protein2_fasta = '/omic/eureka/bioemu/input/villin_headpiece.fasta'
+    complex_name = 'protein_complex'
+    exp_dG = -10.0
+    outdir = '/omic/eureka/bioemu/output'
+    cache_dir = '/tmp/bioemu_cache'
    num_samples = 10
-    batch_size_100 = 10
+    batch_size = 5
+    temperature = 300
+    n_clusters = 5
 }

-// Container configurations
+profiles {
+    standard {
        docker {
            enabled = true
            runOptions = '--gpus all'
        }
+    }

-// Process configurations
+    k8s {
        process {
-    cpus = 1
-    memory = '8 GB'
+            executor = 'k8s'
+        }
+        docker {
+            enabled = true
+        }
    }

-// Execution configurations
-executor {
-    $local {
-        cpus = 4
+    k8s_gpu {
+        process {
+            executor = 'k8s'
+            pod = [[nodeSelector: 'nvidia.com/gpu.present=true']]
+            accelerator = [request: 1, type: 'nvidia.com/gpu']
+        }
+        docker {
+            enabled = true
+        }
+    }
+}
+
+process {
+    container = 'harbor.cluster.omic.ai/omic/bioemu:latest'
+    cpus = 2
    memory = '8 GB'
 }
-}
--- a/params.json
+++ b/params.json
@@ -1,78 +1,81 @@
 {
    "params": {
-        "fasta_list": {
-            "type": "file[]",
-            "description": "FASTA files containing protein sequences",
-            "default": [],
+        "protein1_fasta": {
+            "type": "file",
+            "description": "FASTA file for protein 1",
+            "default": "/omic/eureka/bioemu/input/trp_cage.fasta",
            "required": true,
            "pipeline_io": "input",
-            "var_name": "params.fasta_list",
+            "var_name": "params.protein1_fasta",
            "examples": [
-                ["/omic/olamide/examples/prot1.fasta", "/omic/olamide/examples/prot2.fasta"]
+                "/omic/eureka/bioemu/input/trp_cage.fasta"
            ],
            "pattern": ".*\\.fasta$",
            "validation": {},
-            "notes": "Select one or more FASTA files with protein sequences"
+            "notes": "FASTA file containing the first protein sequence"
+        },
+        "protein2_fasta": {
+            "type": "file",
+            "description": "FASTA file for protein 2",
+            "default": "/omic/eureka/bioemu/input/villin_headpiece.fasta",
+            "required": true,
+            "pipeline_io": "input",
+            "var_name": "params.protein2_fasta",
+            "examples": [
+                "/omic/eureka/bioemu/input/villin_headpiece.fasta"
+            ],
+            "pattern": ".*\\.fasta$",
+            "validation": {},
+            "notes": "FASTA file containing the second protein sequence"
        },
        "outdir": {
            "type": "folder",
-            "description": "Output Directory",
-            "default": "/omic/olamide/output",
+            "description": "Output directory for results",
+            "default": "/omic/eureka/bioemu/output",
            "required": true,
            "pipeline_io": "output",
            "var_name": "params.outdir",
            "examples": [
-                "/omic/olamide/output"
+                "/omic/eureka/bioemu/output"
            ],
            "pattern": ".*",
            "validation": {},
-            "notes": "Select where to save your analysis results"
+            "notes": "Directory where prediction results will be stored"
+        },
+        "complex_name": {
+            "type": "string",
+            "description": "Name for this protein complex",
+            "default": "protein_complex",
+            "required": false,
+            "pipeline_io": "parameter",
+            "var_name": "params.complex_name",
+            "examples": [
+                "protein_complex"
+            ]
        },
        "num_samples": {
            "type": "integer",
            "description": "Number of protein structure samples",
            "default": 10,
-            "required": true,
-            "pipeline_io": "parameter",
-            "var_name": "params.num_samples",
-            "examples": [
-                "10"
-            ],
-            "pattern": "^\\d+$",
-            "validation": {
-                "min": 1
-            },
-            "notes": "More samples provide better coverage of conformational space"
-        },
-        "batch_size_100": {
-            "type": "integer",
-            "description": "Batch size parameter",
-            "default": 10,
            "required": false,
            "pipeline_io": "parameter",
-            "var_name": "params.batch_size_100",
-            "hidden": true,
-            "examples": [
-                "10"
-            ],
-            "pattern": "^\\d+$"
+            "var_name": "params.num_samples",
+            "examples": ["10"],
+            "pattern": "^\\d+$",
+            "validation": {"min": 1},
+            "notes": "More samples provide better coverage of conformational space"
        },
        "temperature": {
            "type": "integer",
-            "description": "Temperature (K) for free energy",
+            "description": "Temperature (K) for free energy calculations",
            "default": 300,
            "required": false,
            "pipeline_io": "parameter",
            "var_name": "params.temperature",
-            "examples": [
-                "300"
-            ],
+            "examples": ["300"],
            "pattern": "^\\d+$",
-            "validation": {
-                "min": 200,
-                "max": 500
-            },
-            "notes": "Temperature in Kelvin for free energy calculations"
+            "validation": {"min": 200, "max": 500},
+            "notes": "Temperature in Kelvin"
        },
        "n_clusters": {
            "type": "integer",
@@ -81,26 +84,19 @@
            "required": false,
            "pipeline_io": "parameter",
            "var_name": "params.n_clusters",
-            "examples": [
-                "5"
-            ],
+            "examples": ["5"],
            "pattern": "^\\d+$",
-            "validation": {
-                "min": 2
+            "validation": {"min": 2}
        },
-            "notes": "Number of clusters for free energy analysis"
-        },
-        "cache_dir": {
-            "type": "folder",
-            "description": "Embeddings cache directory",
-            "default": "/tmp/bioemu_cache",
+        "exp_dG": {
+            "type": "number",
+            "description": "Experimental binding free energy (kcal/mol)",
+            "default": -10.0,
            "required": false,
            "pipeline_io": "parameter",
-            "var_name": "params.cache_dir",
-            "hidden": true,
-            "examples": [
-                "/tmp/bioemu_cache"
-            ]
+            "var_name": "params.exp_dG",
+            "examples": ["-10.0", "-13.1"],
+            "notes": "Reference experimental value for comparison"
        }
    }
 }