slimr: Using R to integrate data and tailor-made population genomic simulations over space and time

Mathematical modelling and simulation is a critically important cornerstone of population genetic practice. At a fundamental level, empirical datasets demand analytical tool-kits that can accommodate their high complexity, and recent developments in sophisticated simulation software have the potential to provide insight into how basic evolutionary processes interact in complex scenarios. However, the most sophisticated simulation software is standalone and has limited integration with software designed for working with empirical data. For example, SLiM >3.0, which is one of the most powerful of the population genomics simulation frameworks for explicitly linking evolutionary dynamics with ecological patterns and processes, has a unique scripting language which limits its availability to many potential users.

In this workshop we will introduce slimr, a new development package for the R statistical programming environment designed to create a seamless link between SLiM and the R development environment, with its powerful data manipulation and analysis tools.

We will show how slimr facilitates smooth integration between genetic and ecological data and simulation in a single software environment. The combination allows the creation of pipelines that begin with data reading, cleaning, and manipulation, proceeds to the construction of empirically-based parameters and initial conditions for simulations, then to the running of numerical simulations, and finally to the retrieval of simulation results in a format suitable for comparisons with the empirical data – aided by advanced analysis and visualization tools provided by R (such as ABC and deep learning).

The workshop will include an introduction followed by a live tutorial, showing how slimr can be used to incorporate population genomic simulations into an empirical workflow, and presentations of two example projects that have already used slimr successfully.

Target Audience

Students or researchers interested in incorporating population genomics simulation into their research programs. Bring along your ideas and questions - anything you've ever wondered about in population genetics that you could learn about through simulation. 

Prerequisites

  • A basic understanding / interest in population genomics.
  • Some familiarity with R would be beneficial.

Computing requirements and delivery mode

The workshop will be run using Rstudio Cloud, and so participants will need a computer with an internet connection in order to follow the tutorial live, and will need to sign up for a (free) Rstudio Cloud account. Invitations to join an Rstudio Cloud workspace will be sent out once the participant list is finalised. Participants will not need to install SLiM or slimr on their own computer for the workshop.

It is possible to participate in the workshop face-to-face or online (a link will be sent after registration). However, depending on the number of online participants, the ability to respond to questions might be limited. All sessions will be recorded for later use.

Morning and afternoon tea will be provided, but participants will need to organise their own lunch at UC campus.

Tentative workshop schedule

9:00am

Introduction and Getting Started tutorial

Russell Dinnage

10:30am

Morning tea break

(provided)

11:00am

Presentation: Effective Population Size Estimation - Insights from Simulations

Bernd Gruber

11:30am

Presentation: What Simulations Can Reveal about Linked Selection

 

Luis Mijangos

12:00pm

Lunch break

 

(bring own or purchase on campus)

1:00pm

Advanced tutorial: How far can we go with slimr? Fully spatially explicit simulations and fitting data to simulations using Approximate Bayesian Computation.

Russell Dinnage

2:00pm

Afternoon tea break

(provided)

2:30pm

Casual Coding Session

 

3:30pm

Workshop end