emilmalta / whisper

Simulate and visualize genetic drift with Wright Fisher

whisper

Simulate and visualise allele trajectories, using Wright-Fisher simulation.

Installation

You can install the development version of whisper from GitHub with:

# install.packages("devtools")
devtools::install_github("emilmalta/whisper")

Load the contents into R environment using

library(whisper)

wf_sim()

The wf_sim function produves a vector of allele frequencies using WF simulation. By default, population size is 250, initial allele frequency is 0.5, and 100 generations are produced.

wf_sim()
#>   [1] 0.500 0.488 0.502 0.520 0.516 0.522 0.534 0.580 0.600 0.596 0.560 0.582
#>  [13] 0.580 0.606 0.616 0.618 0.640 0.640 0.664 0.656 0.660 0.654 0.624 0.638
#>  [25] 0.646 0.624 0.632 0.626 0.604 0.620 0.632 0.618 0.622 0.638 0.640 0.638
#>  [37] 0.664 0.662 0.640 0.592 0.604 0.610 0.634 0.604 0.592 0.566 0.580 0.596
#>  [49] 0.572 0.612 0.600 0.580 0.600 0.588 0.604 0.620 0.646 0.648 0.676 0.696
#>  [61] 0.686 0.668 0.672 0.666 0.648 0.672 0.666 0.698 0.722 0.730 0.762 0.738
#>  [73] 0.742 0.764 0.754 0.760 0.734 0.760 0.796 0.824 0.838 0.812 0.792 0.802
#>  [85] 0.792 0.768 0.784 0.758 0.762 0.776 0.808 0.830 0.858 0.848 0.846 0.860
#>  [97] 0.872 0.890 0.892 0.886

No mutation rates or selection by default. Mutation can be simulated like so:

wf_sim(p = 1/250, mut_to = .8, mut_from = .1)
#>   [1] 0.004 0.816 0.892 0.884 0.894 0.908 0.898 0.904 0.884 0.890 0.882 0.896
#>  [13] 0.898 0.894 0.900 0.896 0.878 0.912 0.894 0.878 0.892 0.892 0.870 0.882
#>  [25] 0.894 0.910 0.888 0.912 0.882 0.902 0.852 0.884 0.872 0.890 0.890 0.888
#>  [37] 0.872 0.860 0.876 0.890 0.888 0.908 0.882 0.868 0.900 0.888 0.886 0.886
#>  [49] 0.902 0.896 0.900 0.888 0.876 0.892 0.878 0.902 0.916 0.896 0.898 0.912
#>  [61] 0.910 0.880 0.856 0.910 0.896 0.894 0.904 0.852 0.866 0.884 0.876 0.882
#>  [73] 0.902 0.896 0.906 0.882 0.888 0.898 0.878 0.890 0.888 0.884 0.876 0.884
#>  [85] 0.922 0.888 0.896 0.902 0.890 0.906 0.890 0.894 0.890 0.886 0.902 0.910
#>  [97] 0.910 0.916 0.908 0.904

Selection assumes a diploid organism, and genotype fitness is provided for each:

# Dominant selection for A:
wf_sim(p = 1/250, fit_aa = .5)
#>   [1] 0.004 0.010 0.022 0.036 0.060 0.102 0.182 0.252 0.374 0.474 0.566 0.578
#>  [13] 0.618 0.730 0.764 0.770 0.804 0.822 0.876 0.878 0.882 0.908 0.914 0.930
#>  [25] 0.928 0.938 0.934 0.946 0.932 0.936 0.932 0.946 0.936 0.924 0.932 0.952
#>  [37] 0.948 0.956 0.956 0.960 0.966 0.980 0.990 0.992 0.986 0.992 0.994 0.998
#>  [49] 0.998 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000
#>  [61] 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000
#>  [73] 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000
#>  [85] 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000
#>  [97] 1.000 1.000 1.000 1.000

wf_plot()

A ggplot of overlaid Wright-Fisher simulations can be made using wf_plot():

wf_plot(num_sims = 100)

Takes same parameters as the wf_sim() function:

wf_plot(p = 1/250, t = 25, fit_Aa = .5, fit_aa = .25) 

wf_shiny()

Run wf_shiny() for an interactive app.