Tidyverse

1 Import and explore with tidyverse

Lets use tidyverse to modify the pizza data.

head(pizza)

Introducing the pipe operator. A key for working in tidyverse. Some fun background it.

pizza %>% select(type)
types_of_pizza <- pizza %>% select(type)
types_of_pizza_char <- as.character(pizza %>% select(type))

Select can also be used to re-name a column or remove a column.

pizza_new <- pizza %>% 
  select(style = type, everything(), -X)

2 filter()

pizza %>% 
  select(style = type, everything(), -X) %>% 
  filter(num_toppings > 2)

Note the change to output the information, rather than save it as a new R object. Also see how the pipe stitches everything together in a tidy way?

pizza %>% 
  select(style = type, everything(), -X) %>% 
  filter(num_toppings > 2) %>% 
  filter(red_sauce == TRUE)

2.0.1 Activity

Try the commented out lines below. Describe the difference.

pizza %>% 
  select(style = type, everything(), -X) %>% 
  # filter(num_toppings > 2 | red_sauce == TRUE) #
  # filter(num_toppings > 2 & red_sauce == TRUE)

3 mutate()

We can also start to introduce math. Mutate adds to the data frame.

This pizza data frame has the cost and the amount of that we want to order. Lets calculate our total order.

pizza$cost
pizza$num_order

pizza %>% 
  select(style = type, everything(), -X) %>% 
  mutate()

4 Wide vs. long format

Create another data frame that is in wide format. For meals and snack throughout the day, how large was each meal and how many dishes does the meal produce?

df_wide <- data.frame(Meal = c('Morning', 'Mid-morning', 'Lunch', 'Mid-afternoon', 'Dinner'),
                 Weight = c(9, 0, 12, 10, 6),
                 Dishes = c(3, 0, 1, 1, 4))
df_wide

df_long <- df_wide %>% 
  pivot_longer(cols = c('Weight', 'Dishes'), names_to = "Attribute", values_to = "Value")

df_long

For long format data, we will have a column for each of the things we are measuring/recording in our data. A column equates to a variable. For wide data, each row has all the information with categories as columns. ‘pivot_longer()’ lengthens the rows of data and decreases the number of rows.

4.1 Activity

Export ‘df_wide’ as a .csv file. Open in Excel and manually change to a long format data frame.

4.2 Convert larger dataset from wide to long

Convert the ASV table to a long format.

head(asv_table)
names(asv_table)


# Compare these two approaches
asv_long <- asv_table %>% 
  pivot_longer(cols = 3:20)

asv_long <- asv_table %>% 
  pivot_longer(starts_with("Gorda"))

Typically data in long format is much more useful for downstream plotting and wrangling.

5 Use tidyverse to compile the ASV data

head(asv_long)
head(metadata)
unique(metadata$SAMPLETYPE)

# Below, pipe in code that will join the metadata with the ASV data. 
## Then isolate samples that were only from the Vent environment. 

# asv_long <- asv_table %>% 
  # pivot_longer(starts_with("Gorda"))

5.1 Activity

Subset the ASV data so that each ASV has greater than or equal to 100 sequences.