assembly-metrics.rmd

---
title: Plot assembly metrics
author: Shaun Jackman
output:
  html_notebook:
    code_folding: hide
  html_document:
    keep_md: true
params:
  input_tsv:
    label: "Input TSV file of assembly metrics"
    value: "dmelanogaster.samtobreak.tsv"
    input: text
---

```{r setup, message=FALSE}
library(dplyr)
library(ggplot2)
library(ggrepel)
library(knitr)
library(readr)
library(scales)
library(tidyr)

knit_print.data.frame <- function(x, ...) kable(x) %>% paste(collapse = "\n") %>% asis_output
input_tsv <- params$input_tsv
```

```{r read-data, message=FALSE}
metrics <- read_tsv(input_tsv) %>%
	mutate(Directory = dirname(File)) %>%
	separate(Directory, c("Trimmer", "Assembler", "k"),
		fill = "left", remove = FALSE) %>%
	mutate(Trimmer = sub("nxtrim", "NxTrim", Trimmer)) %>%
	replace_na(list(Trimmer = "None")) %>%
	mutate(k = as.integer(sub("k", "", k))) %>%
	select(Directory, Trimmer, k,
		starts_with("Scaffold"), starts_with("Contig"), starts_with("Total"),
		everything(), -File, -Assembler)
```

# Scaffold NG50 and NGA50 vs Breakpoints
```{r Scaffold-NGA50-vs-breakpoints}
ggplot(metrics) +
	aes(label = k, x = Total_breakpoints, colour = Trimmer, group = Trimmer) +
	geom_errorbar(aes(ymin = Scaffold_NGA50, ymax = Scaffold_NG50), width = 10) +
	geom_path(aes(y = Scaffold_NGA50), alpha = 0.2) +
	geom_text_repel(aes(y = Scaffold_NG50), segment.color = "#cccccc") +
	scale_x_continuous(name = "Breakpoints", labels = comma) +
	scale_y_continuous(name = "Scaffold NG50 and NGA50", labels = unit_format(unit = "kbp", scale = 1e-3)) +
	scale_colour_brewer(palette = "Set1") +
	expand_limits(y = 0) +
	theme_grey(24)
```

# Contig NGA50 vs breakpoints
```{r Contig-NGA50-vs-breakpoints}
ggplot(metrics) +
	aes(label = k, x = Contig_breakpoints, y = Contig_NGA50, colour = Trimmer, group = Trimmer) +
	geom_path() +
	geom_text_repel(segment.color = "#cccccc") +
	scale_x_continuous(name = "Contig breakpoints", labels = comma) +
	scale_y_continuous(name = "Contig NGA50", labels = unit_format(unit = "kbp", scale = 1e-3)) +
	scale_colour_brewer(palette = "Set1") +
	expand_limits(y = 0) +
	theme_grey(24)
```

# Table of assembly metrics
```{r assembly-metrics-table}
metrics
```