Genomopipe 2.0.0 - Orchestrator, 6 Feedback Loops, Plasmid Design & Desktop GUI

What's New

This is a major release. The core pipeline scripts were substantially rewritten and extended, a complete plasmid design module was added, six feedback loops were implemented, a Python master orchestrator was introduced, and a native Electron desktop GUI was built on top of it all.

Pipeline scripts (Phase 1 - genome_to_design.sh)

    Added dynamic genome_fetch.sh resolution - finds the script via PATH or relative to genome_to_design.sh, with TTY simulation fallback using script
    Added Step 1b: FASTA header sanitization (spaces and pipes → underscores)
    Added Step 2b: repeat masking via RepeatModeler + RepeatMasker; auto-skips for genomes under 1 Mbp; critical for eukaryotes
    Auto-detection of available CPU threads via nproc; GNU parallel used throughout if available
    BRAKER integration now auto-selects OrthoDB partition by querying Entrez taxonomy, downloads and sanitizes the partition FASTA, and repairs GTF transcript_id fields before gffread
    --auto_rnaseq mode: queries SRA, downloads with prefetch + fasterq-dump, aligns with STAR, merges BAMs with samtools
    BRAKER automatically retries in --esmode when fewer than 1000 introns are detected
    Added progress bars and ETA estimates for long-running loops
    RepeatModeler output filtered to suppress per-second countdown noise; full output preserved in logs/repeatmodeler_detail.log
    --force now clears both .done and .failed sentinels
    Organism-change detection: starting a new run automatically if the organism name changes between runs
    Auto-generated README.md run report written at pipeline completion

Phase 2 - plasmid_design_moclo_v3.py (new)

Complete MoClo Golden Gate plasmid design module:

    CDS domestication via DNA Chisel (if installed) with heuristic codon-frequency fallback
    Supports Marillonnet, CIDAR, and JUMP overhang standards with configurable enzyme pairs (BsaI-HFv2 / BpiI defaults)
    Dynamic path resolution: follows latest/ symlink, env vars, or auto-discovers the newest run directory
    Outputs GenBank (.gb) and FASTA files for SnapGene and synthesis ordering

Phase 3 - Six feedback loops (all new)

    FB1 (feedback1_colabfold_to_rfdiffusion.sh): ranks ColabFold predictions by mean pLDDT, feeds top-N structures back into RFdiffusion as fixed motifs, re-runs ProteinMPNN and ColabFold for configurable iterations
    FB2 (feedback2_plddt_mpnn_resample.py): re-runs ProteinMPNN at higher sampling temperature on low-pLDDT backbones; marks unconverged designs after max_iterations
    FB3 (feedback3_blast_to_braker.sh): fetches BLAST hit subject sequences from NCBI, merges with original OrthoDB hints, re-runs BRAKER with the enriched protein hint set, writes annotation diff report
    FB4 (feedback4_domesticated_cds_revalidate.py): extracts domesticated CDS features from plasmid GenBank files, re-folds with ColabFold, flags sequences with pLDDT regression above configurable thresholds
    FB5 (feedback5_designed_proteins_to_annotation.sh): filters validated designs by pLDDT, merges with OrthoDB hints, optionally launches a fresh run on a related organism
    FB6 (feedback6_blast_taxonomy_rerun.py): resolves per-accession taxonomy, assigns each hit to its single most-specific OrthoDB partition keyword, detects and corrects wrong partition selections, re-runs BRAKER with GTF repair and gffread

Master orchestrator - genomopipe.py (new)

    Single entry point for the full pipeline
    YAML, JSON, and plain-text config file support; CLI flags always override config values
    Per-phase sentinel files (.genomopipe_phase1.done, etc.) independent of genome_to_design.sh checkpoints
    --reset clears orchestrator sentinels without touching step-level checkpoints
    Pre-flight script validation before committing to a long run
    Structured summary report (genomopipe_summary.md) written after every run

Desktop GUI - BioForge App (new)

Electron desktop application. See v1.0.0 → v2.0.0 section above for full tab breakdown.
Requirements added since v1.0.0

    Node.js ≥ 18, Electron (npm install in app directory)
    pip install pyyaml pydna (orchestrator + plasmid design)
    pip install dnachisel[reports] (optional - substantially improves domestication)
    braker_env now also requires: repeatmodeler, repeatmasker, star, samtools, sra-tools, gffread

Demo Image