hringbauer/barrierinferpublic

Instructions

This is the source code that was used to simulate and analyze datasets with a barrier; and to visualize the results.

The simulation engine is implemented in grid.py It contains a class Grid.py to simulate data on a grid, and a subclass Secondary_Grid to simulate data on a grid with secondary contact.

The analysis engine is implemented in mle_pairwise.py. The key method, that fits pairwise heterozygosity, is implemented in the class MLE_f_emp.

The Kernel to calculate 2D Diffusion with a barrier is implemented in kernels.py. There, a general Kernel class to calculate covariance matrices between sampling points is implemented, and several subclasses inherit from it. Importantly DiffusionBarrierK0 implements the 2D Diffusion with a barrier.

Importantly, classes to run replicates of various scenarious are implemented in multirun.py. Various subclasses for different scenarios are implemented, with a common framework to simulate and analyze these runs (see bottom of source code). Every subclass there inherits a method to create and analyze a dataset (create_data_set resp. analyze_data_set). A factory method (fac_method) returns these various classes.

The methods from multirun can be run on a cluster, for instance in combination of lines in script_multirun.py and the shell script nickscript.sh.

To visualize the results, figs_paper.py contains various methods to plot data and diagrams.

Also, in analysis.py, several helper methods are implemented, that are called from the other classes (for instance, binning of individual genotype data).

Data Availability

The Scripts used for filtering the hybrid zone data can be found in the subfolder "./SNP Cleaning Scripts"

The filtered Antirrhinum hybrid zone data used for inference (positions and genotypes) can be found in the folder "./DataHZ"