Interpreting genomic data largely relies on simulation software. One important family of software simulates population-level diversity over the course of evolution as a function of selection and demography. The ∂a∂i open-source package, developed by Ryan Gutenkunst (, is the leading software in the field, but is limited when handling large datasets.

This project aims to overhaul the computational engine of the ∂a∂i software by implementing a recently developed approach for solving partial differential equations. The project will use ∂a∂i's interface and code-base, but introduce a new spectrum simulation engine. Using a moment-based approach will improve performance and reduce numerical uncertainty, offering the genetics community with a general-purpose simulation tool that can address larger and more complex datasets than previously possible.

A prototype version of the computational engine is already running. The goal of this project will be to optimize the engine and embed it efficiently in the existing code-base.




  • Simon Gravel