Calculating Diversity

This page will instruct you in the calculation of various common diversity metrics [what are they used for? what do they tell us?]. These steps are intended to precede the calculation of relative abundance.

Alpha Diversity

To calculate alpha diversity, first filter out NAs at the highest level (i.e. superkingdom) and then use the estimate_richness() function from the package phyloseq:

library(phyloseq)

ps.filtered <- ps %>%
  subset_taxa(!is.na(superkingdom))

alphadiv <- estimate_richness(ps.filtered, measures = c("Observed", "Shannon")) %>%
  mutate(barcode_well = rownames(.)) %>%  
  mutate(barcode_well = str_replace_all(barcode_well, "X", "")) %>%  
  mutate(barcode_well = str_replace_all(barcode_well, "\\.", "-")) %>%
  as.data.frame()

Note

Rename ps to something like ps.filtered when filtering out NAs, as shown above. We want to keep the original unfiltered phyloseq available for future steps like calculating relative abundance and performing a clr transform for creating a PCA plot.

This code creates a dataframe with a column barcode_well with the name of each sample and two columns Observed and Shannon with the observed number of taxa and the Shannon diversity respectively of each sample, measures of alpha diversity. We can now join this dataframe to the sample metadata with the following code chunk, based on if a matching barcode_well column exists in your sample metadata:

If barcode_well existsIf barcode_well does not existIf barcode_well exists under a different name

sample_data(ps.filtered) <- sample_data(ps.filtered) %>%
    data.frame() %>%
    left_join(alphadiv, by = "barcode_well")

sample_data(ps.filtered) <- sample_data(ps.filtered) %>%
    data.frame() %>%
    tibble::rownames_to_column("barcode_well") %>%  # Create barcode_well column
    left_join(alphadiv, by = "barcode_well")

sample_data(ps.filtered) <- sample_data(ps.filtered) %>%
    data.frame() %>%
    left_join(alphadiv, by = c("[different name]" = "barcode_well")) # Replace [different name] with the matching column's name

Now that you have a phyloseq object with alpha diversity metrics added to your sample metadata, you can continue with further analyses to analyze the differences in diversity between different samples or groups or continue with calculating relative abundance.