psmelt: Melt phyloseq data object into large data.frame

In joey711/phyloseq: Handling and analysis of high-throughput microbiome census data

Melt phyloseq data object into large data.frame

Description

The psmelt function is a specialized melt function for melting phyloseq objects (instances of the phyloseq class), usually for producing graphics with ggplot2. psmelt relies heavily on the melt and merge functions. The naming conventions used in downstream phyloseq graphics functions have reserved the following variable names that should not be used as the names of sample_variables or taxonomic rank_names. These reserved names are c("Sample", "Abundance", "OTU"). Also, you should not have identical names for sample variables and taxonomic ranks. That is, the intersection of the output of the following two functions sample_variables, rank_names should be an empty vector (e.g. intersect(sample_variables(physeq), rank_names(physeq))). All of these potential name collisions are checked-for and renamed automtically with a warning. However, if you (re)name your variables accordingly ahead of time, it will reduce confusion and eliminate the warnings.

Usage

psmelt(physeq)

Arguments

physeq

(Required). An otu_table-class or phyloseq-class. Function most useful for phyloseq-class.

Details

Note that “melted” phyloseq data is stored much less efficiently, and so RAM storage issues could arise with a smaller dataset (smaller number of samples/OTUs/variables) than one might otherwise expect. For common sizes of graphics-ready datasets, however, this should not be a problem. Because the number of OTU entries has a large effect on the RAM requirement, methods to reduce the number of separate OTU entries – for instance by agglomerating OTUs based on phylogenetic distance using tip_glom – can help alleviate RAM usage problems. This function is made user-accessible for flexibility, but is also used extensively by plot functions in phyloseq.

Value

A data.frame-class table.

See Also

plot_bar

melt

merge

Examples

data("GlobalPatterns")
gp.ch = subset_taxa(GlobalPatterns, Phylum == "Chlamydiae")
mdf = psmelt(gp.ch)
nrow(mdf)
ncol(mdf)
colnames(mdf)
head(rownames(mdf))
# Create a ggplot similar to
library("ggplot2")
p = ggplot(mdf, aes(x=SampleType, y=Abundance, fill=Genus))
p = p + geom_bar(color="black", stat="identity", position="stack")
print(p)

joey711/phyloseq documentation built on Nov. 4, 2022, 1:16 a.m.