In arielebanks/ebanks.package: Fungal OTU analysis

knitr::opts_chunk$set(echo = TRUE)

Processing operational taxonomic unit (OTU) abundance matrices and sample data for downstream analysis

ebanks.package

Once amplicon data is processed, it can be returned as an abundance matrix, data table of OTU taxonomic information, and data table with sample information. The format of these datasets vary depending on the program used to process the amplicon data. The purpose of this package is to turn these separate datasets into one usable taxmap object that can be easily manipulated and analyzed with functions from metacoder, taxa, and phyloseq.

Let's install and load the package.

devtools::install_github("arielebanks/ebanks.package")

library(ebanks.package)

This tutorial will use a sample data table and a fungal OTU abundance matrix. Before you read in the data, make sure tidyverse is loaded.

library(tidyverse)

Now, let's read in and inspect the sample data table.

spartina_sample <- read_csv("../inst/extdata/spartina_sample.csv")

print(spartina_sample)

In this study, leaves, roots, and rhizosphere samples were collected from Spartina alterniflora in oiled and non-oiled sites along the Louisiana Gulf Coast. It displays samples as rows and environmental data and fungal OTUs as columns.

Let's do the same for the abundance matrix.

spartina_otu <- read_csv("../inst/extdata/spartina_otu.csv")

print(spartina_otu)

This abundance matrix is formatted properly with fungal OTUs as rows and samples as columns, but the "taxa_assignment" column is not formatted in a way we can use meaningfully. Also, there is a typo in the first column - "out_id" should read "otu_id".

Parse taxonomic information from abundance matrix

In order to visualize and analyze fungal abundance, richness, and diversity, taxonomic information should be parsed from the abundance matrix. This function takes an OTU dataset and creates a taxmap object with an abundance matrix and parsed taxonomic classifications in two separate tibbles. Additionally, this function will replace "None" with "NAs" in the "taxa_assignment" column and fix the "otu_id" column typo in our spartina_otu dataset.

obj <- parse_tax(spartina_otu)
print(obj)

Our taxmap object says that there are 1328 unique taxa in the dataset, and provides the IDs each was assigned. These IDs can be found in the new column "taxon_id". It also tells us how closely each one will appear in our heat tree. We also see that there are now 2 datasets - obj$data$tax_data and obj$data$class_data.

Manipulating the abundance matrix

Now, we can remove any low abundance counts that may be present due to error, rarefy our counts in case of uneven sampling, and get abundances per taxon and sample tissue type. This function will replace any counts less than 5 with 0 in obj$data$tax_data, and then remove them from the dataset. It will also provide 2 new datasets called obj$data$taxon_abundance and obj$data$tax_sample where abundances per taxon and tissue type can be found. We'll use the taxmap object generated from the previous function as the input here, but we'll name our new object something different so we can compare them.

obj_complete <- fung_abund(obj)
print(obj_complete)

Visualize abundances in all samples using heat tree

Now we can visualize the fungal abundance present in all samples by creating a heat tree from obj_complete.

fung_abund_tree(obj_complete)

arielebanks/ebanks.package documentation built on Dec. 31, 2020, 7:50 p.m.