FUNGuild notebook

Welcome to the FUNGuild tutorial :)

Here are some references: https://rdrr.io/github/brendanf/FUNGuildR/f/README.Rmd https://github.com/UMNFuN/FUNGuild/blob/master/FUNGuild_Manual.pdf

First install FUNGuild

install.packages("devtools")
devtools::install_github("brendanf/FUNGuildR")

load the kitchen sinks…

library(phyloseq)
library(ggplot2)
library(janitor)
library(dplyr)
library(tidyverse)
library(ggpubr)
library(ape)
library(FUNGuildR)
library(splitstackshape)

Then input the FUNGuild database You’ll want to do this every time as the database is always being updated! (This takes a hot second)

fung <- get_funguild_db()

Then lets input our data, We also need to rename one of the rows for FUNGuild to work, It’s just a quirk of the program, FUNGuild wants the taxonomy to be called Taxonomy instead of tax.vector, that DADA2 spits out

fung_data <- read_csv("C:\\Users\\angus\\OneDrive - UNBC\\Angus Ball\\Lab work\\Bioinformatics\\Kenzies Data\\ITS-dada2_nochim_tax.csv")

New names:Rows: 49428 Columns: 91── Column specification ───────────────────────────────────────────────────────────
Delimiter: ","
chr  (2): ...1, tax.vector
dbl (89): CHB1P1, CHB1P10, CHB1P11, CHB1P12, CHB1P13, CHB1P14, CHB1P15, CHB1P2,...
ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

fung_data <- rename(fung_data, Taxonomy = tax.vector)

Then we can assign the taxonomy to the data

fung_guild <- funguild_assign(fung_data, db = fung)

Now it’s time for FUNGuild to become a phyloseq object. See How to make a phyloseq object first, I’ll only be explaining the differences here

#remove the sV value
MetaG_1 <- select(fung_data, -("Taxonomy"))
MetaG_1 <- select(MetaG_1, -(1))
#Remove extraneous samples
MetaG_1 <- select(MetaG_1, -("CHB1P1":"FNP9B3"))

We’ve stolen the the taxonomy table out of the fung_data table, then we’ve split up the single column into each one with respect to taxon

Tax_1 <- select(fung_data, -("CHB1P1":"S4P9PO"))
Tax_1 <- select(Tax_1, -(1))
Tax_1 <- cSplit(Tax_1, "Taxonomy", ":")

Warning: 'as.is' should be specified by the caller; using TRUEWarning: 'as.is' should be specified by the caller; using TRUEWarning: 'as.is' should be specified by the caller; using TRUEWarning: 'as.is' should be specified by the caller; using TRUEWarning: 'as.is' should be specified by the caller; using TRUEWarning: 'as.is' should be specified by the caller; using TRUEWarning: 'as.is' should be specified by the caller; using TRUE

Tax_1 <- rename(Tax_1,
 Kingdom = Taxonomy_1,
 Phylum = Taxonomy_2,
 Class = Taxonomy_3,
 Order = Taxonomy_4,
 Family = Taxonomy_5,
 Genus = Taxonomy_6,
 Species = Taxonomy_7,
  
  )

Then we take the FUNGuild information out of the funguild table

fung_calls <- select(fung_guild, -("CHB1P1":"Taxonomy"))
fung_calls <- select(fung_calls, -(1))

Now we concatenate the table

Tax_1 <- cbind(Tax_1,fung_calls)

Then everything else is the same as a regular phyloseq object

#call row values OTU
rownames(MetaG_1) <- paste0("OTU", 1:nrow(MetaG_1))

rownames(Tax_1) <- paste0("OTU", 1:nrow(Tax_1))



#now real stuff
#convert these bad boys to matrixes
Tax_1 <- as.matrix(Tax_1)
MetaG_1 <- as.matrix(MetaG_1)



#check
class(Tax_1)

[1] "matrix" "array" 

class(MetaG_1)

[1] "matrix" "array" 

OTU_1 = otu_table(MetaG_1, taxa_are_rows = TRUE)
TAX_1 = tax_table(Tax_1)


#combine that data
physeq = phyloseq(OTU_1, TAX_1)



#You have to be absolutely sure you want to run this next command if you do, see phyloseq objects
#physeq <- tax_glom(physeq, taxrank = rank_names(physeq)[5], NArm = FALSE)


#import sample data
Key <- read_csv("C:\\Users\\angus\\OneDrive - UNBC\\Angus Ball\\Lab work\\Bioinformatics\\Kenzies Data\\KeyDiv.csv")

Rows: 61 Columns: 5── Column specification ───────────────────────────────────────────────────────────
Delimiter: ","
chr (5): Name, Location, Layer, Treatment, Site
ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

Key <- data.frame(Key[,-1], row.names=Key$Name)
sampledata = sample_data(Key)

physeq_Key = merge_phyloseq(physeq, sampledata)

Did something break? try reruning “rownames(Tax_1) <- paste0(”OTU”, 1:nrow(Tax_1))“, sometimes its on the fritz