Bioconductor Treesummarizedexperiment

TreeSummarizedExperiment

Dependencies & Environment

Package-intrinsic requirements from the Bioconductor landing page — reproduce in any R environment.

• Version: 2.20.0 · Bioconductor: 3.23 · R: ≥ 4.6
• Depends: SingleCellExperiment, S4Vectors, Biostrings
• Imports: BiocGenerics, ape, rlang, dplyr, SummarizedExperiment, BiocParallel, IRanges, treeio
• System requirements: URL
• Install: BiocManager::install("TreeSummarizedExperiment")

When to Use

• Storing rectangular experimental data alongside hierarchical tree structures using the TreeSummarizedExperiment class.
• Aggregating data to different taxonomic levels (e.g., phylum or class) using aggTSE.
• Subsetting data by specific tree nodes or leaves using subsetByNode.
• Storing reference sequence data per feature using the referenceSeq slot.

When NOT to Use

• For standard single-cell RNA-seq without hierarchical relationships, use SingleCellExperiment because the tree-related slots and overhead are unnecessary.
• For purely manipulating or visualizing phylogenetic trees without rectangular assay data, use ape or ggtree directly because they are specialized for tree operations.

Data Requirements

• An assays matrix representing observed data (e.g., counts) with rows as entities and columns as samples.
• rowData and colData data frames for feature and sample annotations.
• Hierarchical structures provided as phylo objects for rowTree and/or colTree.
• Link information mapping assay rows/columns to tree nodes via rowNodeLab or colNodeLab.
• Optional reference sequences as DNAStringSet or DNAStringSetList.

Key Parameters

• rowTree: A phylo object representing the hierarchical structure on the rows of the assays.
• rowNodeLab: A character vector linking the rows of the assays to the node labels of the rowTree.
• colLevel: The desired aggregation level for columns in aggTSE, specified via node label or node number.
• rowFun: The aggregate function (e.g., sum) applied to the row dimension in aggTSE.
• only.leaf (TRUE): Logical in findDescendant to specify if only leaf descendants should be returned.
• rowFirst (FALSE): Determines the aggregation order in aggTSE when aggregating both dimensions.
• colDataCols: Specifies which columns of colData to keep in the final output of aggTSE.

Best Practices

• Use toTree to convert a taxonomic data.frame into a phylo object before adding it to the object.
• Use changeTree to replace an existing tree and update the mapping if nodes are labeled differently.
• Use aggTSE with colDataCols to speed up aggregation by dropping irrelevant column data.
• Avoid modifying rowLinks or colLinks manually to prevent breaking the link between assays and trees.

Common Pitfalls

• Missing node labels: Row or column names in the assay do not match the node labels of the tree, causing them to be removed with warnings. Fix: Provide the correct mapping via rowNodeLab or colNodeLab during construction.
• Broken links after pruning: Subsetting a tree with ape::keep.tip changes node numbers and breaks links. Fix: Use trackNode to track alias labels and update the LinkDataFrame accordingly.
• Failed tree replacement: Replacing a tree directly with rowTree<- fails if names cannot be matched. Fix: Use changeTree with rowNodeLab when labels differ between the object and the new tree.

Alternatives

• SingleCellExperiment: For storing single-cell data without hierarchical tree structures.
• ape: For general phylogenetic tree manipulation without associated rectangular assay data.
• ggtree: For visualizing phylogenetic trees, which TreeSummarizedExperiment relies on for plotting rather than implementing itself.

Citations

• Huang R, et al. (2021). TreeSummarizedExperiment: a S4 class for data with tree structures. F1000Research.
• Lun and Risso (2020). SingleCellExperiment.

References

• Homepage: https://bioconductor.org/packages/TreeSummarizedExperiment
• Vignette: https://bioconductor.org/packages/release/bioc/vignettes/TreeSummarizedExperiment/inst/doc/Introduction_to_treeSE.html

Run this on BioMate

This skill is the knowledge layer — when, why, and how to use treesummarizedexperiment. To run this analysis on your own data with managed compute, automated QC, and reproducible outputs, use BioMate — free to start.

▶ Open treesummarizedexperiment on BioMate →