Special Does[0]Compute? Session

Dr. Kevin Thornton of the University of California, Irvine has agreed to talk about fwdpy11, the Python package for forward simulation that he has been working on. This provides a means of using Python to interact with simulation input and output, with C++ code doing the heavy lifting.

Link to GitHub page: https://github.com/molpopgen/fwdpy11

This tool benefits from interaction with Kevin’s venerable libsequence package of DNA sequence polymorphism analysis toolkit. Kevin said libsequence is slowly moving toward a version 2.0.

Expected preparation

To prepare for this session, please look at the following:

  1. Take a look at scikit-allel documentation. Kevin will make comparisons to scikit-allel. You can find more information on scikit-allel and analysis of genetic variation at this link.
  2. Have some awareness of Numpy, a Python extension. Please take a look at the tutorial linked here.
  3. Read Kevin’s recent manuscript reporting improved means of “recording and processing tree sequences.” This is important to efficient forward simulations. Here is the link to the Kelleher J, Thornton K, Ashander J, and Ralph P 2018 bioRxiv paper.

In addition, below are two PDFs Kevin sent that are worth looking at.

  1. Overview of what libsequence 2.0 will/could look like in terms of its capabilities for analyzing variation data. The specific examples here have to do with processing the results from Jerome Kelleher’s msprime. See here for link to PDF.
  2. Mini-intro to fwdpy11. The intent here is to create something like a “lab notebook” that a student may show to a PI. “Here are the simulations that I intend and I’m showing here that they seem to work and give the expected results.” The point is that fwdpy11 integrates completely into a work flow that can use any part of Python’s “data science” stack (numpy, pandas, scipy, etc.) and graphical capabilities.

The fwdpy11 intro ends with an example of using a plugin, which means something written in C++ and developed separately from fwdpy11, but using its machinery. Jeremy VanCleve contributed the code shown here, which is an implementation of a “snowdrift” model of social evolution.

See here for link to fwdpy11 mini-intro to fwdpy11.

Note: If there are issues with accessing any of the content for this session, please email any of the following lab members: Chaochih (liux1299@umn.edu), Li Lei (llei@umn.edu), or Peter Morrell (pmorrell@umn.edu).