This repository contains the code for running Expectation-Propagation-ABC for a Neutral evolutionary model.
We present a new model-based simulation approach to study evolution and demography from whole-genome data. It incorporates both coalescent and forward simulations and parallel Expectation Propagation ABC (EP-ABC). EP-ABC makes use of recent approaches in machine-learning and a “divide to conquer” approach to enable efficient distributed computation in genomic analysis. It uses a a population size history model where we allow effective population size fluctuations through time.
There is a Template folder that contains all the files needed to run the EP-ABC and an a CreateInfiles_onepop.sh. The user needs to change the parameters on this file and run it locally to generate the infiles.
It runs in the PBS Pro job scheduler for HPC. Changes need to be done to the bash files to converting to them Slurm.