R Graphics Exercise (Solutions)

library(tidyverse)

Registered S3 methods overwritten by 'ggplot2':
  method         from
  [.quosures     rlang
  c.quosures     rlang
  print.quosures rlang
── Attaching packages ─────────────────────────────────────── tidyverse 1.2.1 ──
✔ ggplot2 3.1.1     ✔ purrr   0.3.2
✔ tibble  2.1.2     ✔ dplyr   0.8.1
✔ tidyr   0.8.3     ✔ stringr 1.4.0
✔ readr   1.3.1     ✔ forcats 0.4.0
── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
✖ dplyr::filter() masks stats::filter()
✖ dplyr::lag()    masks stats::lag()

options(repr.plot.width=4, repr.plot.height=3)

data <- read_tsv('data/gene_counts_raw.txt')

Parsed with column specification:
cols(
  .default = col_double(),
  Label = col_character(),
  Media = col_character(),
  Strain = col_character()
)
See spec(...) for full column specifications.

0. The Label column has 3 pieces of information Sample, Method andd Person in a single cell. Fix this and save the tidy DataFrame as df.

data[1:3, 1:5]

A tibble: 3 × 5

Label	Media	Strain	gene0	gene1
<chr>	<chr>	<chr>	<dbl>	<dbl>
1_MA_J	YPD	H99	1	0
1_RZ_J	YPD	H99	0	0
2_MA_C	YPD	H99	0	0

n <- ncol(data)
data[1:3, (n-4):n]

A tibble: 3 × 5

gene995	gene996	gene997	gene998	gene999
<dbl>	<dbl>	<dbl>	<dbl>	<dbl>
27	325	637	591	848
7	227	508	345	641
31	295	737	738	1006

data %>%
separate(Label, sep='_', into=c("Sample", "Method", "Person")) %>%
select(1:5) %>%
head(3)

A tibble: 3 × 5

Sample	Method	Person	Media	Strain
<chr>	<chr>	<chr>	<chr>	<chr>
1	MA	J	YPD	H99
1	RZ	J	YPD	H99
2	MA	C	YPD	H99

data %>%
separate(Label, sep='_', into=c("Sample", "Method", "Person")) -> df

df %>%
select(1:8) %>%
head

A tibble: 6 × 8

Sample	Method	Person	Media	Strain	gene0	gene1	gene10
<chr>	<chr>	<chr>	<chr>	<chr>	<dbl>	<dbl>	<dbl>
1	MA	J	YPD	H99	1	0	13
1	RZ	J	YPD	H99	0	0	14
2	MA	C	YPD	H99	0	0	10
2	RZ	C	YPD	H99	10	0	18
2	TOT	C	YPD	H99	1	0	0
3	MA	J	YPD	H99	0	0	8

1. Plot a scatter plot of gene100 against gene 1001. Color points by the method used. Save the image as a PNG file ‘fig1.png’ in the ‘figs’ folder.

ggplot(df, aes(x=gene100, y=gene1001, color=Method)) +
geom_point()
ggsave('figs/fig1.png')

Saving 6.67 x 6.67 in image

2. Make a boxplot plot of gene100 counts by method.

ggplot(df, aes(x=Method, y=gene100)) +
geom_boxplot()
ggsave('figs/fig2.png')

Saving 6.67 x 6.67 in image

3. Make a jitter plot of gene100 counts by Media and color the points by method. Set the jitter width to be 0.2.

ggplot(df, aes(x=Media, y=gene100, color=Method)) +
geom_jitter(width=0.2)
ggsave('figs/fig3.png')

Saving 6.67 x 6.67 in image

4. Make a grid of histograms of counts for gene100, with rows showing the person and columns showing the method used.

ggplot(df, aes(x=gene100)) +
facet_grid(Person ~ Method) +
geom_histogram(binwidth=50)
ggsave('figs/fig4.png')

Saving 6.67 x 6.67 in image

5. Make a row of boxplots of log counts of the top 5 genes where each column shows a different method.

Warning: This involves quite a bit of data processing.

genes.top5 <- df %>%
select(starts_with('gene')) %>%
summarize_all(mean) %>%
gather() %>%
arrange(desc(value)) %>%
head(5)

genes.top5

A tibble: 5 × 2

key	value
<chr>	<dbl>
gene1139	884531.49
gene1136	484816.47
gene7418	132744.71
gene363	62325.08
gene7387	45986.86

genes.top5$key

1. 'gene1139'
2. 'gene1136'
3. 'gene7418'
4. 'gene363'
5. 'gene7387'

df %>%
select(c('Method', genes.top5$key)) %>%
gather(gene, count, -Method) %>%
mutate(logcount = log(count)) %>%
ggplot(aes(x=gene, y=logcount)) +
geom_boxplot() +
facet_wrap(~ Method) +
theme(axis.text.x = element_text(angle = 90, hjust = 1))
ggsave('figs/fig5.png')

Saving 6.67 x 6.67 in image