Package 'aggregateBioVar'

Aggregate subject-level biological variation

Description

Given an input gene-by-cell count matrix from a SingleCellExperiment object, sum within-subject gene counts into an aggregate gene-by-subject count matrix. Column metadata accessed by colData are collated by subjectMetaData to remove variables with inter-cell intrasubject variation, effectively retaining between-subject variation. The summary operations are performed across all cell types and within each cell type. A list of SummarizedExperiment objects is returned each with aggregate gene-by-subject count matrix and inter-subject metadata.

Usage

aggregateBioVar(scExp, subjectVar, cellVar)

Arguments

scExp	SingleCellExperiment object containing (at minimum) gene counts and column metadata describing sample identifiers and cell types.
subjectVar	Metadata column name assigning biological sample identity to aggregate within-subject feature counts.
cellVar	Metadata column name assigning cell type. Used for aggregating gene-by-subject count matrices by cell type.

Value

List of SummarizedExperiment objects with gene-by-subject count matrices and variable inter-subject column metadata across and within cell types.

Examples

## Aggregate gene-by-subject count matrix and inter-subject metadata
aggregateBioVar(
    scExp=small_airway,
    subjectVar="orig.ident", cellVar="celltype"
)

---

Within cell type gene-by-subject matrices

Description

Given a vector of unique cell types, calculate a gene-by-subject matrix and inter-subject metadata for a differential expression design matrix.

Usage

countsByCell(scExp, subjectVar, cellVar)

Arguments

scExp	SingleCellExperiment object containing (at minimum) gene counts and column metadata describing sample identifiers and cell types.
subjectVar	Metadata column name assigning biological sample identity to aggregate within-subject feature counts.
cellVar	Metadata column name assigning cell type. Used for aggregating gene-by-subject count matrices by cell type.

Value

List of gene-by-subject and design matrices for each cell type.

See Also

summarizedCounts for aggregate counts and metadata SummarizedExperiment object.

Examples

## Return list of `SummarizedExperiments` with gene-by-subject count matrices
## and subject metadata for each unique `SingleCellExperiment` cell type.
countsByCell(
    scExp=small_airway,
    subjectVar="orig.ident", cellVar="celltype"
)

---

Gene-by-subject count matrix

Description

Convert gene-by-cell count matrix to gene-by-subject count matrix. Row sums are calculated for each feature (i.e. gene) across cells by subject.

Usage

countsBySubject(scExp, subjectVar)

Arguments

scExp	SingleCellExperiment object containing (at minimum) gene counts and column metadata describing sample identifiers and cell types.
subjectVar	Metadata column name assigning biological sample identity to aggregate within-subject feature counts.

Value

S4 DataFrame of gene-by-subject count sums.

See Also

scSubjects for subjects values.

Examples

## Return cell count matrix aggregated by subject.
countsBySubject(scExp=small_airway, subjectVar="orig.ident")

---

SingleCellExperiment subject values

Description

Extract unique values from SingleCellExperiment column (i.e. cell) metadata. Used to determine subject and cell type values.

Usage

scSubjects(scExp, ...)

Arguments

scExp	SingleCellExperiment object containing (at minimum) gene counts and column metadata describing sample identifiers and cell types.
...	Named metadata variables for subjects and cell types.

Value

List of character vectors with unique values from SingleCellExperiment column metadata variables.

Examples

## Examples of metadata column variable names.
names(SummarizedExperiment::colData(small_airway))

## Return list of subject and cell type values from experiment metadata.
scSubjects(scExp=small_airway, subjects="orig.ident", cellTypes="celltype")

---

Small Airway SingleCellExperiment object

Description

Small airway epithelium single-cell RNA sequencing data subset combined from 7 porcine individuals. Genotypes represent wildtype (Genotype WT; n=3) and CFTR-knockout subjects (Genotype CFTRKO; n=4) expressing a cystic fibrosis phenotype.

To access cell counts and column metadata:

• counts: SummarizedExperiment::assay(small_airway, "counts")

• metadata: SummarizedExperiment::colData(small_airway)

Cell types include:

• Secretory cell

• Endothelial cell

• Immune cell

Usage

small_airway

Format

A SingleCellExperiment object with feature counts of 1311 genes (rows) from 2687 individual cells (columns). Features were subset by gene ontology annotation "ion transport" with included child terms ("anion transport", "cation transport", etc.). Includes a count matrix accessed by assay and column metadata accessed by colData. Metadata variables are defined as:

orig.ident

Biological sample identifier (i.e. subject)

nCount_RNA

Number of UMIs (i.e. depth/size factor)

nFeature_RNA

Features with non-zero counts (i.e. identified features)

Genotype

Sample Genotype ('CF' == cystic fibrosis)

• WT = non-CF pig

• CFTRKO = CFTR knockout

Sample

Airway region of sample

celltype

Cell type labels

---

Collate metadata variation

Description

Identify single cell experiment metadata variables that are identical within subject (e.g. genotype, treatment, cell line, sample preparation). Effectively excludes metadata variables containing between cell variation. Used as the design matrix for differential expression analysis.

Usage

subjectMetaData(scExp, subjectVar)

Arguments

scExp	SingleCellExperiment object containing (at minimum) gene counts and column metadata describing sample identifiers and cell types.
subjectVar	Metadata column name assigning biological sample identity to aggregate within-subject feature counts.

Value

Tibble data frame of metadata variables without intrasubject variation from single cell experiment metadata. Rows correspond to aggregated cells (i.e. subject / biological replicate) and columns to metadata attribute variables (e.g. genotype, treatment, cell line).

Examples

## Return experiment metadata sans intrasubject variation.
subjectMetaData(scExp=small_airway, subjectVar="orig.ident")

---

Aggregate feature counts and metadata by subject

Description

Given an input sparse count matrix and corresponding column metadata, aggregate gene counts by subject level. Metadata variables with only inter-subject variation are retained; any variables with cell-level variation within a subject are dropped (e.g. feature / RNA count by cell).

Usage

summarizedCounts(scExp, subjectVar)

Arguments

scExp	SingleCellExperiment object containing (at minimum) gene counts and column metadata describing sample identifiers and cell types.
subjectVar	Metadata column name assigning biological sample identity to aggregate within-subject feature counts.

Value

SummarizedExperiment object with feature counts aggregated by subject and summarized inter-subject metadata.

Examples

## Construct SummarizedExperiment object with gene-by-subject count matrix
## and column metadata summarized to exclude intrasubject variation.
## See `SummarizedExperiment` accessor functions `assay()` and `colData()`
## to access the count matrix and column metadata for downstream analyses.
summarizedCounts(scExp=small_airway, subjectVar="orig.ident")

---