Genome simulation: from 10 minutes to a few seconds by rewriting Python in C
Rewrite of a Python simulation of transposable element dynamics in C, with data-oriented design. Groundwork for the TEDyS software and a 2025 publication.
- Client
- LGDP, University of Perpignan
- Period
- October 2018 to October 2020
- Stack
- Python · C · Data-oriented design
- Outcome
- 100-element simulation: from about 10 minutes to a few seconds
Starting point
The Laboratoire Génome et Développement des Plantes models the dynamics of transposable elements in genomes. The first iteration of the simulation, in Python, took about 10 minutes for 100 elements. Every parameter exploration cost hours.
The delivery
Rewrite of the simulation from Python to C, with data-oriented design: memory layouts follow the way the computation traverses them.
- A 100-element simulation goes from about 10 minutes to a few seconds.
- The laboratory can run far wider parameter sweeps.
Today
This work is the groundwork for the TEDyS software and a 2025 publication.