Package 'exploreDE'

Description

The source-agnostic "bring your own data" constructor. x is a DESeqDataSet (its counts, colData, rowData and factors are used), a SummarizedExperiment (every assay under its name, its colData and its rowData), or the counts themselves (a matrix, or a named list of matrices such as list(counts=, normalised=, vst=, tpm=)). Assay and matrix names become layer names: rawData and raw are raw intensities in proteomics, transformedData the analysis-ready ones. DE tables and pathway statistics are supplied as arguments. Only counts + coldata + omics are required; coldata and rowdata, when given, replace those of x.

Usage

build_explore_h5ad(
  x,
  out,
  omics,
  feature_level = NULL,
  coldata = NULL,
  rowdata = NULL,
  de = list(),
  de_columns = NULL,
  rowdata_columns = NULL,
  pathways = list(),
  enrichment_padj_max = 0.1,
  factors = NULL,
  organism = NA_character_,
  ref_build = NA_character_
)

Arguments

Value

The output path, invisibly.

Examples

library(SummarizedExperiment)
# A prolfqua proteomics result: one contrast, nested in rowData
se <- readRDS(system.file("examples/proteomics/SummarizedExperiment.rds", package = "exploreDE"))
de <- as.data.frame(rowData(se)$constrast_B_vs_A)
path <- build_explore_h5ad(
  x = list(raw = assay(se, "rawData"), transformed = assay(se, "transformedData")),
  out = tempfile(fileext = ".h5ad"),
  omics = "proteomics",
  coldata = colData(se),
  rowdata = de,
  factors = "G_",
  de = list(B_vs_A = de)
)
file.exists(path)

# The SummarizedExperiment itself: rawData and transformedData become the
# layers intensities_raw and intensities
se_path <- build_explore_h5ad(
  se,
  out = tempfile(fileext = ".h5ad"),
  omics = "proteomics",
  rowdata = de,
  factors = "G_",
  de = list(B_vs_A = de)
)

# Counts only
counts_only <- build_explore_h5ad(
  assay(se, "rawData"),
  out = tempfile(fileext = ".h5ad"),
  omics = "proteomics",
  coldata = colData(se)
)

Description

ExploreDataset defines the accessor contract that every concrete subclass must implement. It is not instantiable on its own — calling ExploreDataset$new() raises an error. Use new_explore_dataset() to build a concrete instance, or instantiate a subclass directly.

Methods implemented on the base class read only from private$.config and are inherited unchanged by every subclass. Methods that touch the storage backend are virtual (stop("not implemented")) and must be overridden.

Value

An R6 class generator. ⁠$new()⁠ always errors; the concrete subclasses return an object inheriting from ExploreDataset.

Methods

Public methods

• ExploreDataset$new()

• ExploreDataset$omics_class()

• ExploreDataset$data_type()

• ExploreDataset$has_de()

• ExploreDataset$has_enrichment()

• ExploreDataset$organism()

• ExploreDataset$ref_build()

• ExploreDataset$feature_id_kind()

• ExploreDataset$background_expression()

• ExploreDataset$backend_kind()

• ExploreDataset$backend()

• ExploreDataset$samples()

• ExploreDataset$sample_ids()

• ExploreDataset$n_samples()

• ExploreDataset$factor_names()

• ExploreDataset$factor_levels()

• ExploreDataset$primary_grouping()

• ExploreDataset$secondary_grouping()

• ExploreDataset$sample_factor()

• ExploreDataset$features()

• ExploreDataset$feature_ids()

• ExploreDataset$n_features()

• ExploreDataset$feature_display()

• ExploreDataset$assay_names()

• ExploreDataset$has_assay()

• ExploreDataset$assay()

• ExploreDataset$assay_label()

• ExploreDataset$assay_is_log2()

• ExploreDataset$raw_assay_name()

• ExploreDataset$primary_assay_name()

• ExploreDataset$raw_assay()

• ExploreDataset$primary_assay()

• ExploreDataset$contrasts()

• ExploreDataset$primary_contrast()

• ExploreDataset$de()

• ExploreDataset$de_all()

• ExploreDataset$contrast_meta()

• ExploreDataset$contrast_label()

• ExploreDataset$de_display_extras()

• ExploreDataset$has_pvalue()

• ExploreDataset$enrichment_kinds()

• ExploreDataset$enrichment()

• ExploreDataset$enrichment_input()

• ExploreDataset$enrichment_table()

• ExploreDataset$enrichment_filter()

• ExploreDataset$sgrna_to_gene()

• ExploreDataset$proteomics_choices()

• ExploreDataset$all_pathways()

• ExploreDataset$source()

• ExploreDataset$format()

• ExploreDataset$clone()

ExploreDataset$new()

Always errors: the base class is abstract. Use new_explore_dataset() or a concrete subclass.

Usage

ExploreDataset$new(...)

Arguments

ExploreDataset$omics_class()

Omics class from the config: "genomics" or "proteomics".

Usage

ExploreDataset$omics_class()

ExploreDataset$data_type()

Dataset type from the config, e.g. "RNASeq", "CountQC", "CrisprCountQC" or "proteomics".

Usage

ExploreDataset$data_type()

ExploreDataset$has_de()

TRUE when the dataset has at least one contrast.

Usage

ExploreDataset$has_de()

ExploreDataset$has_enrichment()

TRUE when the config lists at least one enrichment kind.

Usage

ExploreDataset$has_enrichment()

ExploreDataset$organism()

Organism from the config, or NA when unknown.

Usage

ExploreDataset$organism()

ExploreDataset$ref_build()

Reference genome build from the config, or NA when unknown.

Usage

ExploreDataset$ref_build()

ExploreDataset$feature_id_kind()

What a feature ID denotes, from the config, e.g. "gene", "isoform", "protein", "peptide", "phospho" or "guide".

Usage

ExploreDataset$feature_id_kind()

ExploreDataset$background_expression()

Low-signal offset added to expression values before taking log2; from the config, default 10.

Usage

ExploreDataset$background_expression()

ExploreDataset$backend_kind()

Short tag naming the storage backend ("se", "h5ad" or "parquet"). Implemented by subclasses.

Usage

ExploreDataset$backend_kind()

ExploreDataset$backend()

The underlying storage object. Implemented by subclasses.

Usage

ExploreDataset$backend()

ExploreDataset$samples()

Sample metadata as a data.frame, one row per sample, rownames are sample IDs. Implemented by subclasses.

Usage

ExploreDataset$samples()

ExploreDataset$sample_ids()

Sample IDs: the rownames of samples().

Usage

ExploreDataset$sample_ids()

ExploreDataset$n_samples()

Number of samples: the number of rows of samples().

Usage

ExploreDataset$n_samples()

ExploreDataset$factor_names()

Names of the experimental factors from the config; empty when there are none.

Usage

ExploreDataset$factor_names()

ExploreDataset$factor_levels()

Levels of one experimental factor, from the config.

Usage

ExploreDataset$factor_levels(factor)

Arguments

ExploreDataset$primary_grouping()

Name of the factor used as the primary sample grouping, from the config.

Usage

ExploreDataset$primary_grouping()

ExploreDataset$secondary_grouping()

Name of the secondary grouping factor, or NA when there is none.

Usage

ExploreDataset$secondary_grouping()

ExploreDataset$sample_factor()

One column of samples(): the values of a factor, one per sample.

Usage

ExploreDataset$sample_factor(factor)

Arguments

ExploreDataset$features()

Feature annotation as a data.frame, one row per feature, rownames are feature IDs, with a display_name column. Implemented by subclasses.

Usage

ExploreDataset$features()

ExploreDataset$feature_ids()

Feature IDs: the rownames of features().

Usage

ExploreDataset$feature_ids()

ExploreDataset$n_features()

Number of features: the number of rows of features().

Usage

ExploreDataset$n_features()

ExploreDataset$feature_display()

Display names of features, named by feature ID.

Usage

ExploreDataset$feature_display(ids = NULL)

Arguments

ExploreDataset$assay_names()

Canonical names of the assays this dataset provides. Implemented by subclasses.

Usage

ExploreDataset$assay_names()

ExploreDataset$has_assay()

TRUE when name is one of assay_names().

Usage

ExploreDataset$has_assay(name)

Arguments

ExploreDataset$assay()

One assay as a features x samples matrix, optionally subset. Implemented by subclasses.

Usage

ExploreDataset$assay(name, features = NULL, samples = NULL)

Arguments

ExploreDataset$assay_label()

Human-readable label for an assay, from the config; name itself when no label is registered.

Usage

ExploreDataset$assay_label(name)

Arguments

ExploreDataset$assay_is_log2()

TRUE when an assay is on the log2 scale (log2 counts, VST, transformed intensities), from the config. Errors on an assay the dataset does not have.

Usage

ExploreDataset$assay_is_log2(name)

Arguments

ExploreDataset$raw_assay_name()

Canonical name of the raw assay, from the config.

Usage

ExploreDataset$raw_assay_name()

ExploreDataset$primary_assay_name()

Canonical name of the primary assay, from the config.

Usage

ExploreDataset$primary_assay_name()

ExploreDataset$raw_assay()

The raw assay: shortcut for assay(raw_assay_name(), ...).

Usage

ExploreDataset$raw_assay(...)

Arguments

ExploreDataset$primary_assay()

The primary assay: shortcut for assay(primary_assay_name(), ...).

Usage

ExploreDataset$primary_assay(...)

Arguments

ExploreDataset$contrasts()

Names of the DE contrasts, from the config; empty when the dataset has no DE results.

Usage

ExploreDataset$contrasts()

ExploreDataset$primary_contrast()

Name of the primary contrast, from the config; NA when there are no contrasts.

Usage

ExploreDataset$primary_contrast()

ExploreDataset$de()

DE result table for one contrast, as a data.frame. Implemented by subclasses.

Usage

ExploreDataset$de(contrast)

Arguments

ExploreDataset$de_all()

DE result tables of all contrasts, as a list named by contrast.

Usage

ExploreDataset$de_all()

ExploreDataset$contrast_meta()

Metadata list of one contrast from the config: test and reference group, factor, method, thresholds. Errors on an unknown contrast.

Usage

ExploreDataset$contrast_meta(contrast)

Arguments

ExploreDataset$contrast_label()

Display label of a contrast: currently the contrast name itself. Errors on an unknown contrast.

Usage

ExploreDataset$contrast_label(contrast)

Arguments

ExploreDataset$de_display_extras()

Names of extra DE columns shown for display only (e.g. SaintScore), from the config; empty by default.

Usage

ExploreDataset$de_display_extras()

ExploreDataset$has_pvalue()

TRUE when the DE table of at least one contrast has a non-missing raw pValue; FALSE for FDR-only producers such as SAINTexpress.

Usage

ExploreDataset$has_pvalue()

ExploreDataset$enrichment_kinds()

Enrichment kinds available, from the config: any of "ora" and "gsea".

Usage

ExploreDataset$enrichment_kinds()

ExploreDataset$enrichment()

Enrichment results of one contrast and kind, as a flat named list of clusterProfiler results. Implemented by subclasses.

Usage

ExploreDataset$enrichment(contrast, kind)

Arguments

ExploreDataset$enrichment_input()

Inputs for enrichment of one contrast: list(seq_anno, ranked_l2fc, selections). Implemented by subclasses.

Usage

ExploreDataset$enrichment_input(contrast)

Arguments

ExploreDataset$enrichment_table()

Enrichment results of one contrast and kind as one tidy data.frame, derived from enrichment(); an empty data.frame when there are none.

Usage

ExploreDataset$enrichment_table(contrast, kind)

Arguments

ExploreDataset$enrichment_filter()

The p.adjust cut-off applied to the stored enrichment tables, from the config; NA when none was applied.

Usage

ExploreDataset$enrichment_filter()

ExploreDataset$sgrna_to_gene()

Named character vector mapping sgRNA IDs to gene names (CRISPR datasets); NULL otherwise.

Usage

ExploreDataset$sgrna_to_gene()

ExploreDataset$proteomics_choices()

Choices and lookup for the proteomics feature picker, list(choices, lookup); NULL when the dataset has none.

Usage

ExploreDataset$proteomics_choices(kind = c("gene", "protein"))

Arguments

ExploreDataset$all_pathways()

GO terms named in the ⁠GO BP⁠, ⁠GO MF⁠ and ⁠GO CC⁠ columns of features(), as "GO:ID term name" strings; empty when those columns are absent.

Usage

ExploreDataset$all_pathways()

ExploreDataset$source()

Provenance list from the config: backend kind, schema version and build time.

Usage

ExploreDataset$source()

ExploreDataset$format()

Summary of the dataset (type, dimensions, factors, contrasts, assays, enrichment), one line per element. print() on the object shows these lines.

Usage

ExploreDataset$format(...)

Arguments

ExploreDataset$clone()

The objects of this class are cloneable with this method.

Usage

ExploreDataset$clone(deep = FALSE)

Arguments

Examples

se <- readRDS(system.file("examples/proteomics/SummarizedExperiment.rds", package = "exploreDE"))
eds <- new_explore_dataset(se)
inherits(eds, "ExploreDataset")
eds$contrasts()
head(eds$de(eds$primary_contrast()))

# The base class itself cannot be instantiated
try(ExploreDataset$new())

Description

Reads a canonical exploreDE/1.0 AnnData and implements the accessor contract inherited from ExploreDataset. Accepts a path to a .h5ad file or a live anndataR object. This is the backend for user-supplied data ("bring your own data") and the eventual single backend once producers emit canonical .h5ad directly.

Value

An R6 class generator; ⁠$new()⁠ returns an ExploreDatasetH5AD object.

Super class

ExploreDataset -> ExploreDatasetH5AD

Methods

Public methods

• ExploreDatasetH5AD$new()

• ExploreDatasetH5AD$backend()

• ExploreDatasetH5AD$backend_kind()

• ExploreDatasetH5AD$samples()

• ExploreDatasetH5AD$features()

• ExploreDatasetH5AD$de()

• ExploreDatasetH5AD$assay_names()

• ExploreDatasetH5AD$assay()

• ExploreDatasetH5AD$enrichment()

• ExploreDatasetH5AD$enrichment_table()

• ExploreDatasetH5AD$enrichment_input()

• ExploreDatasetH5AD$clone()

ExploreDatasetH5AD$new()

Read a canonical .h5ad or take a live AnnData, build the config from it, and prepare the feature and DE tables.

Usage

ExploreDatasetH5AD$new(input)

Arguments

ExploreDatasetH5AD$backend()

The AnnData parts read at construction, as list(obs, var, layers, varm, uns) of plain R objects.

Usage

ExploreDatasetH5AD$backend()

ExploreDatasetH5AD$backend_kind()

Returns "h5ad".

Usage

ExploreDatasetH5AD$backend_kind()

ExploreDatasetH5AD$samples()

Sample metadata from obs, with the factor_ prefix stripped from column names and a Name column added when missing.

Usage

ExploreDatasetH5AD$samples()

ExploreDatasetH5AD$features()

Feature annotation from var, with feature_id and display_name columns, ordered by the p-value of the primary contrast.

Usage

ExploreDatasetH5AD$features()

ExploreDatasetH5AD$de()

DE table of one contrast from varm, with the legacy aliases log2Ratio, pValue and fdr added; an empty data.frame when the slot is missing. Errors on an unknown contrast.

Usage

ExploreDatasetH5AD$de(contrast)

Arguments

ExploreDatasetH5AD$assay_names()

Canonical assay names from the config's assay map whose layer exists; computed assays are always included.

Usage

ExploreDatasetH5AD$assay_names()

ExploreDatasetH5AD$assay()

One assay as a features x samples matrix (the layer transposed); normalised, normalised_log2, vst, fpkm and tpm are computed when the config maps them to no layer.

Usage

ExploreDatasetH5AD$assay(name, features = NULL, samples = NULL)

Arguments

ExploreDatasetH5AD$enrichment()

Enrichment results of one contrast and kind, rebuilt from uns$enrichment and uns$enrichment_support.

Usage

ExploreDatasetH5AD$enrichment(contrast, kind)

Arguments

ExploreDatasetH5AD$enrichment_table()

Tidy enrichment table of one contrast and kind: the table stored in uns$enrichment when present, otherwise derived from enrichment() as in the base class.

Usage

ExploreDatasetH5AD$enrichment_table(contrast, kind)

Arguments

ExploreDatasetH5AD$enrichment_input()

Enrichment inputs rebuilt from the DE table: features(), log2 fold changes sorted decreasing, and up/down feature selections at the contrast's FDR and log2FC thresholds.

Usage

ExploreDatasetH5AD$enrichment_input(contrast)

Arguments

ExploreDatasetH5AD$clone()

The objects of this class are cloneable with this method.

Usage

ExploreDatasetH5AD$clone(deep = FALSE)

Arguments

Examples

library(SummarizedExperiment)
se <- readRDS(system.file("examples/proteomics/SummarizedExperiment.rds", package = "exploreDE"))
de <- as.data.frame(rowData(se)$constrast_B_vs_A)
path <- tempfile(fileext = ".h5ad")
build_explore_h5ad(
  x = list(raw = assay(se, "rawData"), transformed = assay(se, "transformedData")),
  out = path,
  omics = "proteomics",
  coldata = colData(se),
  rowdata = de,
  factors = "G_",
  de = list(B_vs_A = de)
)
eds <- ExploreDatasetH5AD$new(path)
eds
head(eds$de("B_vs_A"))

Description

ExploreDatasetSE holds the SummarizedExperiment backend reference and implements the accessor contract uniformly for every SE shape. Concrete subclasses (ExploreDatasetSE_RNASeq, ⁠_CountQC⁠, ⁠_CrisprCountQC⁠, ⁠_Proteomics⁠) override two private methods only:

• .build_config(se) — returns the per-instance config list at ⁠$initialize()⁠ time

• .prep() — populates private$.features_df, private$.de_dfs, private$.fid_lookup from the SE

This class is not instantiable on its own — instantiate a concrete subclass instead.

Value

An R6 class generator; the concrete subclasses' ⁠$new()⁠ returns an object inheriting from ExploreDatasetSE.

Super class

ExploreDataset -> ExploreDatasetSE

Methods

Public methods

• ExploreDatasetSE$new()

• ExploreDatasetSE$backend()

• ExploreDatasetSE$backend_kind()

• ExploreDatasetSE$samples()

• ExploreDatasetSE$features()

• ExploreDatasetSE$de()

• ExploreDatasetSE$assay_names()

• ExploreDatasetSE$assay()

• ExploreDatasetSE$enrichment()

• ExploreDatasetSE$enrichment_input()

• ExploreDatasetSE$clone()

ExploreDatasetSE$new()

Wrap a SummarizedExperiment: store it, build the config from it, and prepare the feature and DE tables.

Usage

ExploreDatasetSE$new(se)

Arguments

ExploreDatasetSE$backend()

The wrapped SummarizedExperiment.

Usage

ExploreDatasetSE$backend()

ExploreDatasetSE$backend_kind()

Returns "se".

Usage

ExploreDatasetSE$backend_kind()

ExploreDatasetSE$samples()

Sample metadata from colData(), with the SUSHI ⁠ [Factor]⁠ suffix stripped from column names, rownames set to sample IDs, and a Name column added.

Usage

ExploreDatasetSE$samples()

ExploreDatasetSE$features()

Feature annotation table, prepared from rowData() by the subclass.

Usage

ExploreDatasetSE$features()

ExploreDatasetSE$de()

DE table of one contrast, prepared from rowData() by the subclass; an empty data.frame when none was prepared. Errors on an unknown contrast.

Usage

ExploreDatasetSE$de(contrast)

Arguments

ExploreDatasetSE$assay_names()

Canonical assay names from the config's assay map whose SE assay exists; computed assays are always included.

Usage

ExploreDatasetSE$assay_names()

ExploreDatasetSE$assay()

One assay as a features x samples matrix with rows renamed to feature IDs; normalised_log2, vst, fpkm and tpm are computed when the config maps them to no SE assay.

Usage

ExploreDatasetSE$assay(name, features = NULL, samples = NULL)

Arguments

ExploreDatasetSE$enrichment()

Enrichment results of one kind, read from metadata(se) and flattened to a named list; an empty list when absent. contrast is not used.

Usage

ExploreDatasetSE$enrichment(contrast, kind)

Arguments

ExploreDatasetSE$enrichment_input()

Enrichment inputs read from metadata(se)$enrichInput, with NULL fields and no selections when absent. contrast is not used.

Usage

ExploreDatasetSE$enrichment_input(contrast)

Arguments

ExploreDatasetSE$clone()

The objects of this class are cloneable with this method.

Usage

ExploreDatasetSE$clone(deep = FALSE)

Arguments

Examples

se <- readRDS(system.file("examples/proteomics/SummarizedExperiment.rds", package = "exploreDE"))
eds <- new_explore_dataset(se)
inherits(eds, "ExploreDatasetSE")
eds$assay_names()
dim(eds$raw_assay())

Description

Top-level rowData(se) carries feature annotations; there are no DE results. May include a presentFlag / isPresent assay used as the QC denominator.

Details

Inherits the accessor contract from ExploreDataset via ExploreDatasetSE; overrides only .build_config() and .prep().

Value

An R6 class generator; ⁠$new()⁠ returns an ExploreDatasetSE_CountQC object.

Super classes

ExploreDataset -> ExploreDatasetSE -> ExploreDatasetSE_CountQC

Methods

Public methods

• ExploreDatasetSE_CountQC$clone()

ExploreDatasetSE_CountQC$clone()

The objects of this class are cloneable with this method.

Usage

ExploreDatasetSE_CountQC$clone(deep = FALSE)

Arguments

Examples

library(SummarizedExperiment)
de_result <- readRDS(system.file("examples/genomics/deResult.rds", package = "exploreDE"))
# A count table without DE results, as SUSHI's CountQC writes it
se <- SummarizedExperiment(
  assays = list(counts = assay(de_result, "counts")),
  colData = colData(de_result),
  rowData = rowData(de_result)[, c("gene_id", "gene_name", "description")]
)
eds <- ExploreDatasetSE_CountQC$new(se)
eds
eds$has_de()

Description

Identical to ExploreDatasetSE_CountQC plus a sgRNA → gene mapping parsed from the SE feature IDs, which are assumed to follow the ⁠GENE_<n>_<SEQUENCE>⁠ convention.

Details

Inherits .prep() from ExploreDatasetSE_CountQC unchanged.

Value

An R6 class generator; ⁠$new()⁠ returns an ExploreDatasetSE_CrisprCountQC object.

Super classes

ExploreDataset -> ExploreDatasetSE -> ExploreDatasetSE_CountQC -> ExploreDatasetSE_CrisprCountQC

Methods

Public methods

• ExploreDatasetSE_CrisprCountQC$clone()

ExploreDatasetSE_CrisprCountQC$clone()

The objects of this class are cloneable with this method.

Usage

ExploreDatasetSE_CrisprCountQC$clone(deep = FALSE)

Arguments

Examples

library(SummarizedExperiment)
# Guide counts: 4 guides for each of 10 genes, named GENE_<n>_<SEQUENCE>
set.seed(1)
guides <- sprintf("GENE%d_%d_ACGTACGTACGT", rep(1:10, each = 4), rep(1:4, 10))
counts <- matrix(rpois(40 * 4, 50), nrow = 40, dimnames = list(guides, paste0("S", 1:4)))
samples <- S4Vectors::DataFrame(
  `Condition [Factor]` = rep(c("ctrl", "treated"), each = 2),
  row.names = colnames(counts),
  check.names = FALSE
)
se <- SummarizedExperiment(assays = list(counts = counts), colData = samples)
eds <- ExploreDatasetSE_CrisprCountQC$new(se)
eds
head(eds$sgrna_to_gene())

Description

Reads from a prolfqua-derived SE. Top-level rowData(se) is sparse; feature metadata and per-contrast DE statistics live nested inside ⁠rowData(se)$constrast_<name>⁠ DataFrames (note the SUSHI typo is preserved). The first contrast is the source of truth for feature metadata. feature_id is built from protein_Id (+ site / peptide_Id to disambiguate), and the assay matrices' rownames are remapped via .fid_lookup since SE rownames do not match the built feature_ids.

Details

Inherits the accessor contract from ExploreDataset via ExploreDatasetSE; overrides .build_config() and .prep().

Value

An R6 class generator; ⁠$new()⁠ returns an ExploreDatasetSE_Proteomics object.

Super classes

ExploreDataset -> ExploreDatasetSE -> ExploreDatasetSE_Proteomics

Methods

Public methods

• ExploreDatasetSE_Proteomics$clone()

ExploreDatasetSE_Proteomics$clone()

The objects of this class are cloneable with this method.

Usage

ExploreDatasetSE_Proteomics$clone(deep = FALSE)

Arguments

Examples

se <- readRDS(system.file("examples/proteomics/SummarizedExperiment.rds", package = "exploreDE"))
eds <- ExploreDatasetSE_Proteomics$new(se)
eds
head(eds$feature_ids())
head(eds$de(eds$primary_contrast()))

Description

Reads from a SUSHI ezApp DESeq2 / edgeR / limma result. Top-level rowData(se) carries feature annotations plus inline DE columns for the single contrast described by param$comparison. Optional pre-computed enrichResult / gseaResult objects live in metadata(se).

Details

Inherits the accessor contract from ExploreDataset via ExploreDatasetSE; overrides only .build_config() and .prep().

Value

An R6 class generator; ⁠$new()⁠ returns an ExploreDatasetSE_RNASeq object.

Super classes

ExploreDataset -> ExploreDatasetSE -> ExploreDatasetSE_RNASeq

Methods

Public methods

• ExploreDatasetSE_RNASeq$clone()

ExploreDatasetSE_RNASeq$clone()

The objects of this class are cloneable with this method.

Usage

ExploreDatasetSE_RNASeq$clone(deep = FALSE)

Arguments

Examples

se <- readRDS(system.file("examples/genomics/deResult.rds", package = "exploreDE"))
eds <- ExploreDatasetSE_RNASeq$new(se)
eds
eds$contrasts()
head(eds$de(eds$primary_contrast()))
eds$enrichment_kinds()

Description

Opens the app locally. With no file, it starts on the landing page, where you can upload one or more .h5ad files built with build_explore_h5ad() or load the demo dataset. With file, it opens straight into that dataset; with several files, the first opens in the tabs and the contrasts of the others join it in the Multi DE tabs.

Usage

exploreDE(file = NULL, ...)

Arguments

Details

A SummarizedExperiment opens as new_explore_dataset() classifies it: a SUSHI RNA-seq or prolfqua proteomics result with all its tabs, any other as a Count QC dataset. To see the DE tabs for another SummarizedExperiment, build an .h5ad from it with build_explore_h5ad() and open that.

Value

Called to run the app; returns the value passed to shiny::stopApp() when the app stops (usually NULL), as shiny::runApp() does.

Examples

if (interactive()) {
  exploreDE()
  exploreDE("my_data.h5ad")
  exploreDE(c("design1.h5ad", "design2.h5ad"))
  se <- readRDS(system.file("examples/proteomics/SummarizedExperiment.rds", package = "exploreDE"))
  exploreDE(se)
}

Description

Factory function: loads the input (or accepts a live SummarizedExperiment), classifies its shape via detect_se_shape(), and returns the appropriate concrete subclass.

Usage

new_explore_dataset(input)

Arguments

Details

Accepted inputs:

• A live SummarizedExperiment object.

• A character path to an .rds / .qs2 file.

• A character path to a .RData file (Mageck EzResult convention).

• A character path to a directory containing one of the standard SUSHI artefacts (deResult.rds, rawData_final.rds, etc.).

• An ⁠http(s)://⁠ URL pointing at an .rds file.

Value

An object inheriting from ExploreDataset.

Examples

# From a live SummarizedExperiment
se <- readRDS(system.file("examples/proteomics/SummarizedExperiment.rds", package = "exploreDE"))
eds <- new_explore_dataset(se)
class(eds)[1]

# From an .h5ad file written by build_explore_h5ad()
path <- build_explore_h5ad(
  SummarizedExperiment::assay(se, "rawData"),
  out = tempfile(fileext = ".h5ad"),
  omics = "proteomics",
  coldata = SummarizedExperiment::colData(se)
)
eds_h5ad <- new_explore_dataset(path)
class(eds_h5ad)[1]
eds_h5ad$has_de()

Description

Checks subclass-agnostic invariants by walking the accessor API:

Usage

validate_explore_dataset(eds, strict = TRUE)

Arguments

Details

1. The runtime class is a concrete ExploreDataset subclass — not the abstract base or intermediate.

2. Cross-references resolve: factor_names() are columns of samples(); every contrasts() entry has a working de() with the effect_log2fc and score_fdr columns; primary_grouping() is a member of factor_names(); primary_contrast() is in contrasts() (or NA).

3. Assays: raw_assay_name() and primary_assay_name() are in assay_names(), every assay has a log2 fact (assay_is_log2()), and the raw assay's rownames / colnames agree with feature_ids() / sample_ids().

Cost is proportional to the number of contrasts (one de(c) round-trip each) plus a small constant of other accessor calls.

Value

Invisibly, TRUE on success (strict mode); or list(ok = logical(1), errors = character) (non-strict mode).

Examples

se <- readRDS(system.file("examples/proteomics/SummarizedExperiment.rds", package = "exploreDE"))
eds <- new_explore_dataset(se)
validate_explore_dataset(eds)
validate_explore_dataset(eds, strict = FALSE)