nextflow.config

/*
========================================================================================
    nf-proteindesign Nextflow config file
========================================================================================
    Default config options for all compute environments
----------------------------------------------------------------------------------------
*/

// Nextflow plugins
plugins {
    id 'nf-schema@2.1.1'
}

// Global default params, used in configs
params {

    // Input options
    input                      = null
    
    // Test mode: run ONLY the design tool (no downstream ProteinMPNN / Boltz-2 / etc.)
    test_design_only           = false
    
    // ========================================================================
    // Protein design tool selection
    // ========================================================================
    // Choose which generative model drives the design stage.
    //   'boltzgen'        — BoltzGen (original, default)
    //   'complexa'        — Proteina-Complexa (flow-matching approach)
    //   'rfdiffusion_v3'  — RFdiffusion3 (all-atom diffusion, RosettaCommons)
    //
    // Each tool uses its own samplesheet columns and parameters (see below).
    // Everything downstream (ProteinMPNN, Boltz-2, IPSAE, etc.) is shared.
    // ========================================================================
    protein_design_tool        = 'boltzgen'          // 'boltzgen', 'complexa', or 'rfdiffusion_v3'

    // ========================================================================
    // BoltzGen design parameters (used when protein_design_tool = 'boltzgen')
    // ========================================================================
    // Samplesheet columns: sample_id, design_yaml, target_sequence,
    //   structure_files (optional), protocol, num_designs, budget,
    //   reuse (optional), target_msa (optional), target_template (optional)
    // ========================================================================
    cache_dir                  = null                // Cache directory for BoltzGen model weights (~6GB), defaults to ~/.cache
    boltzgen_config            = null                // Optional: Path to custom BoltzGen config YAML to override defaults
    steps                      = null                // Optional: Comma-separated list of steps to run (e.g., 'filtering')

    // ========================================================================
    // Proteina-Complexa design parameters (used when protein_design_tool = 'complexa')
    // ========================================================================
    // Samplesheet columns: sample_id, target_pdb, pipeline_config,
    //   target_sequence, target_msa (optional), target_template (optional)
    // ========================================================================
    complexa_ckpt_dir          = null                // Path to Complexa checkpoint directory (required for GPU inference)
    complexa_container         = '307946633589.dkr.ecr.eu-west-2.amazonaws.com/rashmi/proteina-complexa:latest'
    complexa_search_algorithm  = 'best-of-n'         // Search algorithm: best-of-n, single-pass, beam-search, fk-steering, mcts
    complexa_nsteps            = 400                  // Number of diffusion sampling steps (generation.args.nsteps)
    complexa_replicas          = 2                    // Number of replicas for best-of-n search (generation.search.best_of_n.replicas)
    complexa_batch_size        = 16                   // Dataloader batch size for generation (generation.dataloader.batch_size)
    complexa_extra_args        = ''                   // Additional Hydra overrides (e.g., '++seed=42 ++generation.args.guidance_w=2.0')

    // ========================================================================
    // RFdiffusion v3 design parameters (used when protein_design_tool = 'rfdiffusion_v3')
    // ========================================================================
    // Samplesheet columns: sample_id, design_yaml, structure_files,
    //   num_designs, budget, target_msa (optional), target_sequence (optional),
    //   target_template (optional)
    // ========================================================================
    rfdiffusion_v3_ckpt_dir    = null                // Path to RFdiffusion3 checkpoint dir (auto-downloaded if null)
    rfdiffusion_v3_container   = 'rosettacommons/foundry:latest'  // Container with rfd3 CLI
    
    // ProteinMPNN sequence optimization options
    run_proteinmpnn            = true                // Enable ProteinMPNN sequence optimization of designed structures (set to false to disable)
    mpnn_sampling_temp         = 0.1                 // Sampling temperature (0.1-0.3 recommended, lower = more conservative)
    mpnn_num_seq_per_target    = 8                   // Number of sequence variants to generate per structure
    mpnn_batch_size            = 1                   // Batch size for ProteinMPNN inference
    mpnn_seed                  = 37                  // Random seed for reproducibility 
    mpnn_backbone_noise        = 0.02                // Backbone noise level (0.02-0.20, lower = more faithful to input)
    mpnn_save_score            = true                // Save per-residue scores
    mpnn_save_probs            = false               // Save per-residue probabilities (large files, use for detailed analysis)
    mpnn_fixed_chains          = null                // Chains to keep fixed (e.g., 'A,B' - typically the target chains)
    mpnn_designed_chains       = null                // Chains to design (e.g., 'C' - typically the binder chain)
    
    // Boltz-2 structure prediction options (for refolding ProteinMPNN sequences)
    run_boltz2_refold          = true                // Enable Boltz-2 structure prediction for ProteinMPNN sequences (set to false to disable)
    boltz2_cache               = null                // Cache directory for Boltz-2 model weights (~6GB), defaults to ~/.boltz
    boltz2_num_recycling       = 3                   // Number of recycling steps (3-5 recommended)
    boltz2_num_diffusion       = 200                 // Number of diffusion samples (higher = slower but more accurate)
    boltz2_use_msa             = false               // Use MSA server for template search (slower but more accurate)
    boltz2_predict_affinity    = true                // Predict binding affinity (log IC50 values)
    boltz2_torch_precision     = 'medium'            // Torch float32 matmul precision: 'medium', 'high', or 'highest' (for Tensor Cores)

    // IPSAE scoring options
    run_ipsae                  = true                // Enable IPSAE scoring of Boltz-2 structures (set to false to disable)
    ipsae_pae_cutoff           = 10                  // PAE cutoff for IPSAE calculation (Angstroms, default: 10)
    ipsae_dist_cutoff          = 10                  // Distance cutoff for CA-CA contacts (Angstroms, default: 10)
    
    // PRODIGY binding affinity prediction options
    run_prodigy                = true                // Enable PRODIGY binding affinity prediction on final designs (set to false to disable)
    prodigy_selection          = null                // Chain selection for PRODIGY (e.g., 'A,B'). If null, auto-detects from structure
    
    // Foldseek structural search options
    run_foldseek               = true                // Enable Foldseek structural similarity search (requires foldseek_database, set to false to disable)
    foldseek_database          = null                // Path to Foldseek database directory (e.g., AlphaFold/Swiss-Model, required if run_foldseek is true)
    foldseek_database_name     = 'afdb'              // Database name within the directory (e.g., 'afdb', 'swissprot', 'pdb')
    foldseek_evalue            = 0.001               // E-value threshold for reporting matches (lower = more stringent)
    foldseek_max_seqs          = 100                 // Maximum number of target sequences to report
    foldseek_sensitivity       = 9.5                 // Search sensitivity (1.0-9.5, higher = more sensitive but slower)
    foldseek_coverage          = 0.0                 // Minimum fraction of aligned residues (0.0-1.0, higher = more global alignment)
    foldseek_alignment_type    = 2                   // Alignment type: 0=3Di only, 1=TMalign (global), 2=3Di+AA (local, default)
    
    // Metrics consolidation and reporting options
    run_consolidation          = true                // Enable consolidated metrics report generation (set to false to disable)
    report_top_n               = 10                  // Number of top designs to highlight in report
    
    // Output options
    outdir                     = './results'
    publish_dir_mode           = 'copy'
    
    // Max resource options
    // Defaults only, expecting to be overwritten
    max_memory                 = '128.GB'
    max_cpus                   = 16
    max_time                   = '240.h'
    max_gpus                   = 1                       // Maximum GPUs per process (increase for multi-GPU systems)
    
    // GPU acceleration options
    // NOTE: The following processes support GPU acceleration:
    //   - PROTEINA_COMPLEXA_DESIGN: Requires GPU for flow-matching inference
    //   - PROTEINMPNN_OPTIMIZE: Optional GPU support, accelerates sequence optimization
    //   - BOLTZ2_REFOLD: Requires GPU for structure prediction / refolding
    //   - FOLDSEEK_SEARCH: Optional GPU support, provides 4-27x speedup for structure searches
    // Ensure your compute environment has NVIDIA GPUs and Docker/Singularity GPU support enabled

    // Boilerplate options
    tracedir                   = "${params.outdir}/pipeline_info"
    help                       = false
    version                    = false
    validate_params            = true
    show_hidden_params         = false
    
}

// Load base.config by default for all pipelines
includeConfig 'conf/base.config'

// Container options
docker.enabled             = false
podman.enabled             = false

// Set default container registry
apptainer.registry         = ''
docker.registry            = 'docker.io'

// Capture exit codes from upstream processes when piping
process.shell = ['/bin/bash', '-euo', 'pipefail']

manifest {
    name            = 'seqeralabs/nf-proteindesign'
    author          = 'Florian Wuennemann'
    homePage        = 'https://github.com/seqeralabs/nf-proteindesign'
    description     = 'Nextflow pipeline for protein design with BoltzGen, Proteina-Complexa, or RFdiffusion v3 and parallel sample processing'
    mainScript      = 'main.nf'
    nextflowVersion = '!>=23.04.0'
    version         = '1.0.0'
}

// Load config profiles
profiles {
    debug {
        dumpHashes             = true
        process.beforeScript   = 'echo $HOSTNAME'
        cleanup                = false
    }
    
    docker {
        docker.enabled         = true
        podman.enabled         = false
        docker.runOptions = '--gpus all'
    }
    
    singularity {
        singularity.enabled    = true
        singularity.autoMounts = true
        // Enable GPU support for Singularity/Apptainer
        singularity.runOptions = '--nv'
    }
    
    apptainer {
        apptainer.enabled      = true
        apptainer.autoMounts   = true
        // Enable GPU support for Apptainer
        apptainer.runOptions   = '--nv'
    }
        
    test_design_nanobody {
        includeConfig 'conf/test_design_nanobody.config'
    }
    
    test_design_peptide {
        includeConfig 'conf/test_design_peptide.config'
    }
    
    test_design_protein {
        includeConfig 'conf/test_design_protein.config'
    }
    
    test_design_rfdiffusion_v3 {
        includeConfig 'conf/test_design_rfdiffusion_v3.config'
    }
    
    test_design_proteina_complexa {
        includeConfig 'conf/test_design_proteina_complexa.config'
    }
}