cohort_relabund: Creates a relative abundance cohort plot
In EESI/theseus: Analysis and Visualization Tools for Microbial Community Data

Description Usage Arguments Details Value See Also Examples

This function plots the relative abundance of taxa within a phyloseq object 'PS' according to thier pre-determined cohort memberships. See Details for more information.

cohort_relabund(PS, xvar = "SampleID", taxfill = "Phylum", comp1, comp2,
  comp1lab = c("Decreased Comp1", "No Change Comp1", "Increased Comp1"),
  comp2lab = c("Decreased Comp2", "No Change Comp2", "Increased Comp2"),
  justdata = FALSE, PSisRelAbund = FALSE)

`PS`	(required) A phyloseq object.
`xvar`	(required) Variable in sample_data(PS) to be displayed on the x-axis. Defaults to 'SampleID'.
`taxfill`	Taxonomic level to display in plot. Defaults to 'Phylum'.
`comp1`	(required) First comparison (hence 'comp1') object of 'DESeqResults' class, or a dataframe with similar structure. In the case that a 'DESeqResults' class object is not being used, the object must contain a 'log2FoldChange' vector/column. Row names must be a subset (but not necessarily a proper subset) of taxa_names(PS). See Details for more information.
`comp2`	(required) Second comparison (hence 'comp2') object. Refer to documentation for 'comp1' for remaining details.
`comp1lab`	Labels for comparison 1. Defaults to c('Decreased Comp1', 'No Change Comp1','Increased Comp1').
`comp2lab`	Labels for comparison 2. Defaults to c('Decreased Comp2', 'No Change Comp2','Increased Comp2').
`justdata`	Return only the data table (no plot). Defaults to FALSE.
`PSisRelAbund`	Does the PS object contain compositional (relative abundance) taxa counts? Defaults to FALSE.

General

The results from a single pairwise comparison, such as pre- and post-treatment, carried out with the DESeq2 package can be plotted or read and interpreted in tabular form with relative ease. When two pairwise comparisons are being performed, interpreting the results becomes more difficult. This function is intended to assist with interpreting the results from multiple differential abundance analyses carried out with the DESeq2 r-package. This function takes a phyloseq object ('PS') and two 'DESeqResults' objects ('comp1', 'comp2') and plots the relative abundance of taxa within 'PS', partitioning the taxa according to their membership to one the 9 possible cohort combinations determined by their values specified within 'comp1' and 'comp2'.

Approach

The DESeq function carries out differential abundance testing and produces a 'DESeqDataSet' object. The results function can be used to access the results and create a 'DESeqResults' object, which is a subclass of DataFrame. Note that the 'alpha' parameter for results can be used to specify the significance level of the test being performed. NOTE: Testing with DESeq2 must be carried out, and the non-significant taxa should be removed from the 'comp1' and 'comp2' objects before using this function. Using the log2FoldChange columns in 'comp1' and 'comp2' this function identifies which taxa decrease, do not change, or increase over course of both comparisons. Because there are three options for both comparisons there are 3^2=9 possible combinations, or cohorts, which an OTU may fall into. These cohort assignments are used when plotting the relative abundance plot.

A ggplot object.

plot_bar transform_sample_counts DESeq results DESeqDataSet

## Not run: 
cohort_relabund(
  PS=prune_samples(sample_data(WWTP_Impact)$site %in% c(1,2,3,4),
                   WWTP_Impact),
  comp1=sigtab,
  comp2=sigtab.2vs3,
  comp1lab=c('Decreased at Effluent',
             'No change at Effluent',
             'Increased at effluent'),
  comp2lab=c('Decreased btwn plants',
             'No change btwn plants',
             'Increased btwn plants'))

## End(Not run)