SVModeller Documentation

SVModeller is a computational simulator designed to create synthetic human haplotypes containing embedded structural variants (SVs). The simulator has been trained on an extensive catalogue of curated polymorphic SVs identified across 1,019 samples from the 1000 Genomes Project sequenced with Oxford Nanopore long-read technology.

Key Features

  • Realistic SV Embeddings: Includes repetitive insertions such as Variable Number of Tandem Repeats (VNTRs), Mobile Element Insertions (MEIs: Alu, LINE-1, SVA), transductions, and mitochondrial insertions (NUMTs).
  • Structural Variant Types: Supports Alu, LINE-1, SVA, transductions, NUMTs, VNTRs, duplications, inversions, deletions, inverted duplications, orphan insertions, and deletions.
  • Modular Design: 5 independent modules covering distribution extraction, event generation, deletion processing, genome modification, and sub-clonal read simulation.
  • Containerized: Run easily via Docker, Conda, or Python environment.

Documentation Sections

  1. Installation Guide: How to set up SVModeller using Docker, Conda, or Pip.
  2. Module Reference Guide: Comprehensive parameter breakdown, input/output files, and sample commands for Modules 1 through 5.
  3. Running as a Nextflow Pipeline: Guide on executing the end-to-end pipeline.
  4. Data & Reference Sets: Access pre-calculated model distributions and reference datasets on Zenodo.
  5. Interactive Command Builder: Dynamically construct your pipeline or module run commands in the browser.

Workflow Overview

graph TD
    VCF_IN[Input VCF Insertions] --> M1[Module 1 - Distribution Analysis]
    M1 --> DIST[Features and Probabilities TSVs]
    DIST --> M2[Module 2 - Generate Insertion Events]
    REF_FA[Reference Genome FASTA] --> M2
    M2 --> INS_TAB[Insertions Table TSV]

    VCF_DEL[Input VCF Deletions] --> M3[Module 3 - Generate Deletion Events]
    REF_FA --> M3
    M3 --> DEL_TAB[Deletions Table TSV]

    INS_TAB --> M4[Module 4 - Genome Modification]
    DEL_TAB --> M4
    REF_FA --> M4
    M4 --> MOD_FA[Modified Genome FASTA]

    MOD_FA --> M5[Module 5 - Read Simulation and BAM Alignment]
    REF_FA --> M5
    M5 --> BAM[Simulated BAM and FastQ Reads]

Quick Examples

Docker (Single Module)

# Pull the v0.5.0 Docker container
docker pull ghcr.io/biocorecrg/svmodeller:0.5.0

# Run Module 1
docker run --rm -v $(pwd):/data ghcr.io/biocorecrg/svmodeller:0.5.0 \
    Module1.py --file_path /data/VCF_Insertions.vcf --chromosome_length /data/chr_length.txt

Nextflow (End-to-End Pipeline)

# Run the entire pipeline end-to-end with the test dataset
nextflow run . -profile applecontainer -params-file params.test.yaml -resume

Developers & License

Developed by Ismael Vera-Munoz (orcid.org/0009-0009-2860-378X) at the Repetitive DNA Biology (REPBIO) Lab at the Centre for Genomic Regulation (CRG).

Distributed under the AGPL-3.0 License.