continuous_anc.Rmd

---
title: "Continuous Ancestral State"
output: github_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
#devtools::install_github("arborworkflows/aRbor")
library(aRbor)
library(geiger)
library(knitr)
```

## Loading in sample data and running function

```{r}
tr  <- read.tree("/Users/carrietribble/Desktop/arbor_hackathon_docs/heliconia/Heliconia_all_dated_mod.phy")
trait <- read.csv('/Users/carrietribble/Desktop/arbor_hackathon_docs/heliconia/Heliconia_phylogeny_matrix.csv')

td <- treeplyr::make.treedata(tr, trait)

ace_arbor <- aRbor::aceArbor( select_(td, "InflorL"), charType='continuous', aceType = 'marginal')
```

## Summary of Continuous Ancestral State Reconstruction

```{r include = FALSE}
trait_name <- attributes(ace_arbor)$names

orig_mean <- mean(as.matrix(attributes(ace_arbor)$td$dat))

anc_mean <- mean(ace_arbor$InflorL$estimate)

orig_range <- c(min(as.matrix(attributes(ace_arbor)$td$dat)),
                max(as.matrix(attributes(ace_arbor)$td$dat)))

anc_range <- c(min(ace_arbor[[1]]$estimate),
               max(ace_arbor[[1]]$estimate))

anc_range_CI <- c(min(ace_arbor[[1]]$lowerCI95),
               max(ace_arbor[[1]]$upperCI95))

# ordering and binning confidence intervals by age 

CI_age <- data.frame(age = branching.times(attributes(ace_arbor)$td$phy),
                     CI_width = ace_arbor[[1]]$upperCI95 - ace_arbor[[1]]$lowerCI95)
```

The mean value of `r paste(trait_name)` extant values is `r orig_mean`, the reconstructed nodes have a mean value of `r anc_mean`. The original range of your data values is `r orig_range[1]` to `r orig_range[2]`. The estimated node values range from `r anc_range[1]` to `r anc_range[2]`. When we include the 95% confidence invervals for those estimates, the estimates range from `r anc_range_CI[1]` to `r anc_range_CI[2]`. A portion of those estimates are displayed below. 

```{r, echo = FALSE}

if (nrow(ace_arbor[[1]]) > 20) {
  rows_to_include <- 20
} else {rows_to_include <- nrow(ace_arbor[[1]])}

kable(ace_arbor[[1]][1:rows_to_include,])
```


The plot below shows the relationship between the width of your 95% confidence intervals and time as indicated by node ages. 

```{r echo = FALSE}

plot(CI_width~age, data = CI_age,
     pch = 19,
     main = trait_name,
     xlab = "Node Age",
     ylab = "Width of 95% CI for Estimated Values")

```