Automated Reporting for Systems Biology Modeling Workflows

[matthiaskoenig]

Automated Reporting for Systems Biology Modeling Workflows

Background

Computational systems biology relies heavily on standardized modeling and simulation workflows, most notably SBML (Systems Biology Markup Language) for models, SED-ML (Simulation Experiment Description Markup Language) for simulation setups or PEtab for parameter optimization workflows. While robust toolchains exist for parameter estimation and sensitivity analysis, the communication of results remains a major bottleneck.

Currently, results are often scattered across scripts, log files, plots, and tables, making it difficult to ensure reproducibility, transparency, and accessibility for collaborators and the wider community. There is a clear need for automated, reproducible reporting pipelines that transform modeling workflows into human-readable, well-structured reports.

This project builds on and contributes directly to sbmlsim, an open-source Python framework for simulation, parameter optimization, and sensitivity analysis of SBML models:

• https://github.com/matthiaskoenig/sbmlsim
• https://matthiaskoenig.github.io/sbmlsim/

Several existing model repositories already use CI/CD workflows to execute simulations, but reporting is mostly manual or ad hoc. This project addresses this gap by integrating standardized report generation into modeling workflows.

Goal

The goal of this GSoC project is to design and implement an automated reporting framework for systems biology modeling workflows based on SBML, SED-ML, and PETab integrated into sbmlsim.

The project will:

• Generate interactive web reports using Quarto (interactive figures and tables).
• Generate static PDF reports using Typst (publication-ready summaries and figures).
• Integrate report generation into CI/CD pipelines (GitHub Actions).
• Automatically publish web reports via GitHub Pages (github.io).

The final outcome will be reusable reporting modules and templates within sbmlsim that can be adopted across multiple systems biology and systems medicine projects.

Difficulty Level: Medium

This project is rated Medium because it combines Python development, scientific workflows, CI/CD automation, and report generation. While advanced mathematical modeling is not required, familiarity with reproducible computational workflows is important.

Size and Length of Project

• medium: 175 hours
• 12 weeks

Skills

Essential skills

• Python
• Markdown
• Git / GitHub

Nice to have skills

• Quarto
• Typst
• CI/CD (GitHub Actions)
• Systems biology or scientific computing background

Public Repository

Primary development repository:

• https://github.com/matthiaskoenig/sbmlsim
• https://matthiaskoenig.github.io/sbmlsim/

Contributing to sbmlsim on develop branch:

• https://github.com/matthiaskoenig/sbmlsim/tree/develop

Example model repositories using CI/CD workflows:

• https://github.com/matthiaskoenig/canagliflozin-model

All code will be open source and publicly available throughout and after GSoC.

Potential Mentors

• Matthias König — konigmatt@googlemail.com

AI Usage Policy

AI tools (e.g. large language models for code suggestions, refactoring, or drafting documentation) are permitted as assistive tools, under the following conditions:

• The contributor must fully understand, review, and validate all AI-assisted outputs.
• The student remains fully responsible for correctness, licensing compliance, and scientific validity.
• AI tools must not be used to fabricate data, results, or evaluations.
• Substantial AI assistance should be transparently documented in commit messages or pull requests.

The project emphasizes reproducibility, transparency, and responsible use of AI as a supporting instrument—not a substitute for scientific reasoning or software engineering.

[atulbraj]

Hi @matthiaskoenig,

Thank you for posting this project idea! Automated reporting for computational workflows is definitely a pain point I've observed.

Quick clarifying question: Would the reporting framework primarily focus on template-based configuration (Quarto/Typst templates), or would it also involve building programmatic Python APIs for dynamic report generation?

I'm currently exploring GSoC projects in computational systems biology and appreciate the detailed doubt description.

Best regards,
Atul

[matthiaskoenig]

The main focus is on template-based configuration. Since the report-generation functions will have well-defined method signatures, they can be exposed as an API at a later stage (e.g. via FastAPI) with minimal effort. We already rely heavily on Pydantic models and type annotations throughout the codebase, which makes adding an API layer straightforward—though this is explicitly not the primary goal at this stage.

Instead, the emphasis is on establishing a set of reusable building blocks for standard reports, built on clearly defined result data. Examples include:

• Parameter optimization: goodness-of-fit plots, AIC, RMSE, performance on training and validation data, parameter tables, confidence intervals, and profile likelihoods
• Sensitivity analysis: heatmaps, summary tables, and related visualizations

The reports should also document the methods and settings used for their generation, ensuring that they are fully self-contained outputs. This makes them suitable for direct inclusion in publications or for hosting online.

I hope this clarifies the intent.

[dyrpsf]

Hello @matthiaskoenig,

Thank you for sharing this interesting project idea. Automated and reproducible reporting is indeed a critical need in computational systems biology, and I find this project highly relevant.

I have prior experience contributing to SBSCL and JSBML using Java, where I submitted multiple pull requests and received positive feedback. This has given me a strong understanding of standardized biological modeling workflows and the importance of reproducibility and well-structured tooling.

I also have a solid background in Python, Git/GitHub, and scientific programming, and I’m keen to work on integrating Quarto/Typst-based reporting, CI/CD automation, and reusable templates within sbmlsim.

I’m particularly interested in helping design maintainable reporting modules that can generate self-contained, publication-ready outputs and be easily adopted across different projects.

I would be grateful for any guidance on how I can start contributing and prepare myself to be a strong GSoC candidate for this project.

Thank you for your time and consideration.

Best regards,
Deepak

[matthiaskoenig]

@dyrpsf Please send me your CV via email and we will make a short online meeting next week.
Best Matthias

[atulbraj]

Hi @matthiaskoenig,

I'm very interested in this GSoC project! I've created a proof-of-concept to demonstrate my understanding of the requirements.

Prototype:

• Repository: https://github.com/atulbraj/sbmlsim-report-prototype
• Live Report: https://github.com/atulbraj/sbmlsim-report-prototype/blob/main/parameter_optimization_report.html

What it demonstrates:

1. Template-based reporting with Quarto
2. Interactive visualizations (matplotlib/seaborn)
3. Self-contained HTML output with embedded assets
4. Reproducibility documentation
5. Publication-ready formatting
6. Proposed integration strategies with sbmlsim (XResult, UnitsInformation)

I'd love to discuss:

1. How this aligns with your vision for sbmlsim reporting
2. Integration points with the existing codebase
3. Priority order for different report types (parameter optimization, sensitivity analysis, PRISMA)

Would you be available for a brief discussion about the project scope? Let me according to your availabiliy.

Best regards,
Atul