reads_graphs.Rmd

---
title: "Reads_graphs"
author: "Mayher"
date: "8/1/2019"
output: html_document
---

I first install and download all necessary libraries and packages
```{r setup, include=FALSE}

require("dplyr")
require("tidyr")
require(ggplot2)
```

Then, I am reading in my file, and then only keeping the columns that I need: mapped, mapped no mispriming, reads, and the Sample Line ID. 
```{r file}
df <- read.delim("../data/MetaDataSequencing.txt")
keeps <- c("line","fraction", "reads", "mapped", "Mapped_noMP")
frac_dif <- df[keeps]
```

Next, I am dividing the data set into it's two categories: nuclear and total
```{r nuc and total}
nuclear <- subset(frac_dif, fraction == "nuclear", c(fraction, line, reads, mapped, Mapped_noMP))
total <- subset(frac_dif, fraction == "total", c(fraction, line, reads, mapped, Mapped_noMP))
```

Here, I create my data matrices. Each are proportions of reads. There is mapped, mapped with mispriming, mapped without mispriming, and unmapped. I do this for both nuclear and total. 
```{r prop df}
#nuclear proportions
nuc_mapped_prop <- data.matrix(nuclear$mapped/nuclear$reads)
nuc_mapped_noMP_prop <- data.matrix(nuclear$Mapped_noMP/nuclear$reads)
nuc_mapped_MP <- nuc_mapped_prop - nuc_mapped_noMP_prop
nuc_none <- 1 - nuc_mapped_prop
nuc_lines <- data.matrix(nuclear$line)

#total proportions
total_mapped_prop <- data.matrix(total$mapped/total$reads)
total_mapped_noMP_prop <- data.matrix(total$Mapped_noMP/total$reads)
total_mapped_MP <-total_mapped_prop - total_mapped_noMP_prop
total_none <- 1 - total_mapped_prop
total_lines <- data.matrix(total$line)
```

Then, I combine these proportions into a large data frame called "combination", which I can then easily create my plots.The gather() function tidies my data so that it is in the format that ggplot uses to create the bar plots. I then make the type column a factor with 3 levels. These levels allow me to order the stacks in the way I want them to. 
```{r combination}
nuc_combination <- data.frame(nuc_lines, nuc_mapped_noMP_prop, nuc_mapped_MP, nuc_none)
nuc_combination <- gather(nuc_combination, nuc_mapped_noMP_prop, nuc_mapped_MP, nuc_none, key = "type", value = "count")
nuc_combination$type <- factor(nuc_combination$type, levels=c("nuc_none", "nuc_mapped_MP", "nuc_mapped_noMP_prop"))

total_combination <- data.frame(total_lines, total_mapped_noMP_prop, total_mapped_MP, total_none)
total_combination <- gather(total_combination, total_mapped_noMP_prop, total_mapped_MP, total_none, key = "type", value = "count")
total_combination$type <- factor(total_combination$type, levels=c("total_none", "total_mapped_MP", "total_mapped_noMP_prop"))
```


Finally, I create my graphs, one for nuclear and the other for total. geom_col allows me to make bar graphs based off of heights in the data, instead of frequency, which geom_bar does. I stacked the bar plots using three proportions - mapped with mispriming, mapped without mispriming, and unmapped. 
```{r pressure, echo=FALSE}
ggplot(data = nuc_combination, aes(x = nuc_lines, y = count, fill = type)) + geom_col() + theme(axis.text.x = element_text( hjust = 0,vjust = 1, size = 6, angle = 90)) + xlab("lines") + ylab("proportions") + scale_fill_brewer(palette="Dark2", name = "type of read:", labels = c("unmapped", "mapped + MP", "mapped + noMP")) +  ggtitle("nuclear lines vs. proportion")

ggplot(data = total_combination, aes(x = total_lines, y = count, fill = type)) + geom_col() + theme(axis.text.x = element_text( hjust = 0,vjust = 1, size = 6, angle = 90)) + xlab("lines") + ylab("proportions") + scale_fill_brewer(palette="Dark2", name = "type of read:", labels = c("unmapped", "mapped + MP", "mapped + noMP")) +  ggtitle("total lines vs. proportion")


```

Note that the `echo = FALSE` parameter was added to the code chunk to prevent printing of the R code that generated the plot.
```{r}

```