---
title: "Examples"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Examples}
  %\VignetteEngine{knitr::rmarkdown}
  %\usepackage[utf8]{inputenc}
---


```{r echo=FALSE}
library(knitr)
opts_chunk$set(comment = NA, 
               fig.width = 7, 
               fig.height = 5,
               fig.align = "center")
```


## 1. Obtain the neutral distribution of Tajima's D

For 20 samples from a panmictic population:

```{r}
 library(coala)
 model <- coal_model(20, 2000) +
   feat_mutation(2) +
   feat_recombination(1) +
   sumstat_tajimas_d()
 stats <- simulate(model, seed = 15)
 plot(density(stats$tajimas_d, na.rm = TRUE), 
      main = "Neutral Distribution of Tajiam's D")
```



## 2. Generate the site frequency spectrum
For a neutral model with two populations and migration:
```{r}
model2a <- coal_model(c(10, 10), 100) +
   feat_mutation(10) +
   feat_recombination(5) +
   feat_migration(0.5, symmetric = TRUE) +
   sumstat_sfs(population = "all")
stats <- simulate(model2a, seed = 20)
barplot(stats$sfs / sum(stats$sfs), 
        names.arg = seq_along(stats$sfs), 
        col = 3)
```

And again, but now with a bottleneck in one population:
```{r}
model2b <- model2a +
  feat_size_change(0.1, population = 2, time = 0.25) +
  feat_size_change(1, population = 2, time = 0.5)
stats <- simulate(model2b, seed = 25)
barplot(stats$sfs / sum(stats$sfs), 
        names.arg = seq_along(stats$sfs), 
        col = 4)
```



## 3. Plot the nucleotide diversity against the mutation rate:
```{r}
model3 <- coal_model(10, 50) +
  feat_mutation(par_prior("theta", sample.int(100, 1))) +
  sumstat_nucleotide_div()
stats <- simulate(model3, nsim = 40)
mean_pi <- sapply(stats, function(x) mean(x$pi))
theta <- sapply(stats, function(x) x$pars[["theta"]])
plot(theta, mean_pi, pch = 19, col = "orange")
abline(lm(mean_pi ~ theta), col = "red2", lty = 3)
```

If you have a nice example for using coala, feel free to extend this 
vignette via a pull request [on GitHub](https://github.com/statgenlmu/coala)!