Week 8: LSODA Gets Smarter

-

Hi all,

Okay, Continuing our story from last blog. So, after spending the past week chasing down the issue with SBML Test Suite failures (which I elaborated on in my last blog post), I finally got a breakthrough that brought us closer to a robust and reliable LSODA integration in SBSCL.

Revisiting the SBML Test Suite Challenges

Previously, around 12% of the tests were failing and 2% were throwing outright errors. After narrowing the root cause down to event handling discrepancies, I spent this entire week diving deeper into tackling this challenge.

Dr. Dräger shared a fantastic resource (link), which not only confirmed my theory but also revealed that previous developers had faced similar problems while implementing the Rosenbrock solver in SBSCL. It turns out that precise event time detection is a non-trivial challenge when dealing with SBML models.

When events are triggered between timepoints (say at tτ, which lies between tτ−1 and tτ+1), evaluating events only at those boundary points may completely miss the triggering condition. This demands a step-size adaptation mechanism capable of “hitting” the exact event time.

The Rosenbrock solver addressed this by adapting its step size whenever an event alters the computed state vector. The minimal step size ultimately defines how finely events can be captured.

Simulating Event-Aware LSODA Behavior

Taking inspiration from that mechanism, I refactored the computeChange() method of LSODA. Here's what I did:

  • Let Sol(next) be the result from LSODA at time t + h.

  • Process events and rules after computing this result.

  • If processing causes any change in Sol(next), it implies an event was triggered.

  • In that case, reduce h (by a factor of 1/10) and try again.

  • If the solution stabilizes, increase h (by a factor of 6) for the next step.

This loop ensures that event-triggered changes are captured at the most precise point possible.

Results?

After making these changes, I ran the test suite again. The outcome was fantastic. Finally, I saw some actual progress:

  • Over 95% of the test cases passed immediately!

  • A small number still failed, mostly related to fast reaction tags.

  • After further digging and fixing a subtle issue — ~99% of the test cases passed.

This is the result map from the SBMLTestRunner.

For the remaining 1% that still fails, I couldn’t track down the precise cause. However, with this result I'm convinced that the core LSODA integration is now working correctly and is well-aligned with the SBSCL pipeline, including SBML compatibility. As suggested by Taichi, I’ve opened this issue with all the details around the remaining test failures. 

Next Steps

  • Start working on comprehensive documentation for LSODA including internal working, algorithm flow, and best practices.

  • Collaborate with mentors to finalize any loose ends in integration.

Personal Note

Due to a some miscommunication, we couldn't have our weekly meeting. We’ll catch up on Slack and resume discussions next week. And yes — on a personal note, I received my midterm evaluations result and the feedback. I passed! I can’t express how happy and grateful I feel. Thank you Google, NRNB and my wonderful mentors for this incredible opportunity. 

Thanks for reading! ✨
- Bye! 👋

Comments

Post a Comment

Popular posts from this blog

Week 7: Events, Exhaustion and Evaluations

-
Image
Hi all, This week has probably been the toughest one yet. Not just mentally, but also in terms of hours spent chasing bugs that I had no idea about. I’ve logged over 50 hours on the project, and ironically, there’s not much “visible” progress to show. But deep down, I know this is part of the process, and what matters is that I’ve been able to get to the root of a tricky problem, more on that below. Updates From This Week Made the PR changes suggested by Dr. Dräger and Taichi. Ran the SBMLTestSuite using the wrapper program I wrote last week. I had already shared the test results with the team on Slack. Results? Mixed bag: Passed: ~85% of the test cases (Happy somewhat!) Wrong Output: ~12–13% Error: ~2% (No idea, why?) Results Map from SBMLTestRunner So naturally, I spent the entire week narrowing down those failed and wrong output cases. And guess what? I’ve finally figured out what's going wrong in most of the wrong output cases. The Issue: Event Timin...
Read more

My GSoC 2025 Journey Begins

-
Image
Hi,  I am Ayush Baranwal, a second-year engineering undergrad at Jawaharlal Nehru University, New Delhi, India. I am incredibly excited to share that I have been selected for the Google Summer of Code 2025 with the National Resource for Network Biology (NRNB) organization. This summer, I will be working on the project " Developing a highly efficient numerical ODE solver in Systems Biology Simulation Core Library (SBSCL). " I am incredibly thankful to my wonderful mentors, Prof. Dr. Andreas  Dräger , Dr. Akira Funahashi, Taichi Araki and Arthur Neumann, for their constant support and guidance throughout the proposal period. I am truly grateful for the trust they have put in me, and I am fully committed to making this project a success. What This Blog Series is About This blog series will document my journey throughout GSoC. Prof.  Dräger  suggested keeping a public log during our first meeting post-selection—more on that in the next blog! The goal is to keep my work ...
Read more

Week 1: Coding Period Begins

-
Hi everyone, The official coding period has now begun, and I've diving deeper into the LSODA internals. Continuing from the last blog post, I focused primarily on refactoring helper methods within the LSODA package, while also  setting the foundation for a robust and modular design.   After reading some of the core literature on LSODA, I decided to examine the libLSODA codebase—the C library of LSODA translated from its original Fortran implementation in ODEPACK. My first step was to outline the overall program flow of the LSODA algorithm , which helped me visualize where and how different components fit together.   Upon exploring the main lsoda function, I chose to start by implementing and refining the helper functions , as they form the building blocks of the algorithm. I thoroughly reviewed the SBSCL LSODA implementation so far and identified areas for improvement. Here's a more detailed  explanation of my work ( PR ),  followed by wha...
Read more