phyloseq (version 1.16.2)

tip_glom: Agglomerate closely-related taxa using single-linkage clustering.

Description

All tips of the tree separated by a cophenetic distance smaller than h will be agglomerated into one taxa using merge_taxa.

Usage

tip_glom(physeq, h = 0.2, hcfun = agnes, ...)

Arguments

physeq

(Required). A phyloseq-class, containing a phylogenetic tree. Alternatively, a phylogenetic tree phylo will also work.

h

(Optional). Numeric scalar of the height where the tree should be cut. This refers to the tree resulting from hierarchical clustering of cophenetic.phylo(phy_tree(physeq)), not necessarily the original phylogenetic tree, phy_tree(physeq). Default value is 0.2. Note that this argument used to be named speciationMinLength, before this function/method was rewritten.

hcfun

(Optional). A function. The (agglomerative, hierarchical) clustering function to use. Good examples are agnes and hclust. The default is agnes.

...

(Optional). Additional named arguments to pass to hcfun.

Value

• An instance of the phyloseq-class. Or alternatively, a phylo object if the physeq argument was just a tree. In the expected-use case, the number of OTUs will be fewer (see ntaxa), after merging OTUs that are related enough to be called the same OTU.

Details

Can be used to create a non-trivial OTU Table, if a phylogenetic tree is available.

For now, a simple, ``greedy'', single-linkage clustering is used. In future releases it should be possible to specify different clustering approaches available in R, in particular, complete-linkage clustering appears to be used more commonly for OTU clustering applications.

See Also

merge_taxa

agnes

hclust

cophenetic.phylo

phylo

Examples

data("esophagus")
# for speed
esophagus = prune_taxa(taxa_names(esophagus)[1:25], esophagus)
plot_tree(esophagus, label.tips="taxa_names", size="abundance", title="Before tip_glom()")
plot_tree(tip_glom(esophagus, h=0.2), label.tips="taxa_names", size="abundance", title="After tip_glom()")