gbdraw: a genome diagram generator for microbes and organelles

gbdraw is a secure, serverless web application and command-line tool developed to generate customizable circular and linear genome diagrams for microbes and organelles directly within the user's browser, eliminating the need for data uploads while supporting both programmatic batch processing and graphical user interface accessibility.

The Gist

Imagine you have a giant, complex instruction manual for a living machine—like a bacteria or a tiny virus. This manual, called a genome, is just a long string of letters (A, C, T, G) that tells the organism how to build itself.

Scientists need to draw pictures of these manuals to understand how they work, how they compare to each other, and how they evolved. But until now, making these pictures has been a bit of a headache.

Here is the story of gbdraw, a new tool that makes drawing these genome maps easy, safe, and fun for everyone.

The Problem: The "Too Hard" vs. "Too Risky" Dilemma

Before gbdraw, scientists had two main choices, and both had flaws:

1. The "Code-It-Yourself" Route: There were powerful tools, but you had to be a master programmer to use them. It was like trying to build a custom car engine by welding every part yourself. Great if you're an expert, but impossible for a regular biologist who just wants to see a picture.
2. The "Upload-It-All" Route: There were websites where you could upload your data and get a picture back. But this was like sending your top-secret family recipes to a stranger's kitchen to be copied. You had to trust them with your private, unpublished data, and you often had to wait for the internet to process it.

The Solution: gbdraw (The "Magic Sketchbook")

gbdraw is a new tool that fixes both problems. Think of it as a magic sketchbook that lives right inside your web browser.

1. It's a "Serverless" Magic Trick

The coolest thing about gbdraw is that it never sends your data to a remote server.

• The Analogy: Imagine you have a secret recipe. Instead of mailing it to a restaurant to be cooked, you bring your own ingredients and a portable stove to the restaurant. You cook it right there in front of the customers, and the restaurant never touches your food.
• How it works: gbdraw uses a special technology (WebAssembly) that lets your computer do all the heavy lifting inside your browser. Your genome data never leaves your machine. It's 100% private and secure, even if you are using a public library computer.

2. Two Ways to Play: The "Playground" and the "Factory"

gbdraw is designed for two types of people:

• The Playground (Web App): If you just want to see a picture quickly, you go to the website. You drag and drop your file, click a few colorful buttons to change the style, and poof—a beautiful diagram appears. No coding required. It's like using a "Paint" app but for DNA.
• The Factory (Command Line): If you are a computer wizard who needs to make 1,000 pictures automatically, you can use the command-line version. It's like a factory robot that can churn out genome maps while you sleep.

3. What Can It Draw?

gbdraw is incredibly flexible. It can draw:

• Circular Maps: Like a clock face or a pizza, perfect for bacteria and mitochondria (the power plants in our cells).
• Linear Maps: Like a long strip of road, perfect for comparing how different viruses or bacteria are rearranged.
• The "Detective" Feature: It has a built-in engine called LOSAT that acts like a super-fast detective. It can instantly compare two different genomes to see which parts are similar (like finding matching puzzle pieces) without you needing to install complex, heavy software on your computer.

4. Making It Pretty

You can customize everything.

• Colors: It comes with 55 different color palettes (like "Soft Pastel" or "Rainbow").
• Labels: You can tell it exactly what to write on the map. Do you want to see the gene names? The protein names? Or just the big picture? You decide.
• Editing: You can click on the map in your browser and move things around or hide specific parts, just like editing a photo.

Why Does This Matter?

In the past, a biologist might have spent three days writing code or wrestling with complex software just to make one figure for a presentation. With gbdraw, they can do it in three minutes.

It bridges the gap between experts (who need deep control) and everyone else (who just needs a clear picture). Whether you are studying a tiny virus that infects shrimp or the massive DNA of a human mitochondrion, gbdraw helps you visualize the invisible world of life, safely and easily.

In short: gbdraw is the "Canva" of genome diagrams—powerful enough for pros, simple enough for anyone, and secure enough for your most sensitive secrets.

Technical Summary

1. Problem Statement

The paper identifies a significant gap in the current ecosystem of genome visualization tools. Researchers face a "trade-off" between accessibility and functionality:

• Programming Libraries: Tools like Circos, pyGenomeViz, and gggenomes offer high customization and batch processing capabilities but require extensive coding skills and deep parameter understanding.
• Graphical Applications: Desktop tools (e.g., EasyFig, Artemis) and web platforms (e.g., Proksee) offer user-friendly interfaces but often suffer from limitations such as:
  • Data Privacy: Web-based tools typically require uploading sensitive genomic data to remote servers.
  • Installation Complexity: Local tools often require complex dependency management (e.g., BLAST+ suite) and local installation.
  • Topology Rigidity: Most tools are restricted to either circular or linear visualization, forcing users to switch software for different dataset aspects.
  • Comparative Limitations: Generating pairwise comparisons often requires pre-calculated BLAST outputs or local installation of heavy software suites.

2. Methodology and Implementation

gbdraw is a Python-based tool designed to bridge these gaps by offering both a Graphical User Interface (GUI) and a Command-Line Interface (CLI).

• Core Architecture:

  • Language: Python 3 (version 3.10+).
  • Libraries: Utilizes Biopython for parsing GenBank/DDBJ/GFF3/FASTA files, pandas for data handling, and svgwrite for generating SVG graphics.
  • Web Application (Serverless): The web version runs entirely within the user's browser using Pyodide (WebAssembly). This executes the core Python logic locally, eliminating the need for a backend server and ensuring data never leaves the user's machine.
  • Image Generation:
    • CLI: Uses CairoSVG to convert SVG to PNG/PDF.
    • Web: Uses the HTML5 Canvas API for PNG and JavaScript libraries (jsPDF, svg2pdf.js) for PDF conversion.

• Comparative Genomics Engine (LOSAT):

  • Integrates LOSAT, a homology search engine built for the browser.
  • Enables on-the-fly generation of pairwise genomic comparisons (equivalent to TBLASTX or BLASTN) without requiring the user to install the BLAST+ suite.
  • For large-scale tasks where browser performance is limited, users can import pre-calculated BLAST output files (tab/comma-separated).

• Input/Output:

  • Inputs: Self-contained GenBank/DDBJ files, or combinations of GFF3 annotation and FASTA sequence files.
  • Outputs: Circular and linear genome diagrams in SVG, PNG, and PDF formats.
  • State Management: Users can export the entire visualization session (parameters, tracks, edits) as a .json file for reproducibility or sharing.

3. Key Features and Capabilities

• Dual Topology Support: Seamlessly renders both circular (concentric tracks for replicons) and linear (synteny and structural rearrangements) diagrams.
• Visualization Tracks:
  • Annotated features (CDS, tRNA, etc.).
  • GC content and GC skew tracks (useful for identifying origins of replication and genome reshuffling).
  • Pairwise comparison ribbons (linking homologous regions).
• Customization:
  • 55 Built-in Color Palettes: Including specific schemes for organellar genomes (e.g., OGDRAW style).
  • Granular Control: Users can define colors for specific feature types or individual features, adjust label priorities (e.g., preferring 'gene' over 'product'), and use blacklists/whitelists.
  • Interactive Editing: Direct editing of features and labels is possible in the web app; CLI users can apply edits via TSV files.
• Scope: Suitable for microbial genomes, organelles (mitochondria/chloroplasts), and viral genomes. While not designed for whole eukaryotic genomes, it can visualize specific regions of interest and handles features with introns.

4. Results and Demonstrations

The paper demonstrates the tool's versatility through several case studies:

• Circular Diagrams: Successfully visualized the human mitochondrial DNA, tobacco plastid genome, and Vibrio nigripulchritudo (showing six replicons on a single canvas).
• Linear Diagrams:
  • E. coli chromosomes linked by LOSATN pairwise matches.
  • Majaniviruses (endogenous DNA viruses) linked by TBLASTX matches.
  • Lividomycin-like biosynthetic gene clusters connected by ribbons.
  • Betacoronaviruses (SARS-CoV-2, SARS-CoV-1, MERS-CoV) showing comparative alignments.
• Performance: The tool effectively handles the visualization of complex datasets without requiring local software installation or data upload to external servers.

5. Significance and Impact

• Data Security: The serverless architecture is a major advancement for researchers handling unpublished or sensitive genomic data, as processing occurs locally in the browser.
• Accessibility: It democratizes high-quality genome visualization by removing the barrier of coding skills (via the GUI) while maintaining the power for automation (via the CLI).
• Workflow Integration: By removing the dependency on local BLAST installations for comparative analysis, it streamlines the workflow for experimental biologists who may lack bioinformatics infrastructure.
• Availability: The tool is open-source (MIT License), available via the Bioconda channel for easy installation, and accessible as a free web application at https://gbdraw.app/.

In summary, gbdraw fills a critical niche by providing a secure, flexible, and user-friendly solution for microbial and organellar genome visualization, effectively balancing the needs of both computational and experimental biologists.