Package 'seqout'

Description

Uses the accession shape, so no request is made.

Usage

accession_kind(accession, archive = FALSE)

AccessionKind(accession, archive = FALSE)

Arguments

Value

One of "series", "study", "experiment", "sample", "run", "biosample" or "submission", or NA_character_ when the shape is not recognised. With archive = TRUE, a named character vector of kind and archive, the latter one of "geo", "sra", "ena", "ddbj", "gsa", "arrayexpress", "gea", "bioproject" or "biosample".

Examples

AccessionKind("GSE168652")
AccessionKind("SRR13927092")
AccessionKind("not-an-accession")

# PRJCA resolves to GSA under the PRJ shape
AccessionKind("PRJCA042384", archive = TRUE)
AccessionKind("PRJCA042384", archive = TRUE)[["archive"]]

Description

Datasets named by an author's publications.

Usage

author(name, limit = 200, con = .con())

Author(name, limit = 200, con = .con())

Arguments

Value

A tibble of projects.

Examples

Author("Aviv Regev")

Description

join = "inner", the default, keeps the features common to every matrix and warns when that drops any. Peak-by-cell matrices called per sample rarely share a feature space, so the difference can be large.

Usage

bind_counts(
  x,
  labels = NULL,
  max_cells = NULL,
  strict = FALSE,
  join = c("inner", "outer")
)

BindCounts(
  x,
  labels = NULL,
  max_cells = NULL,
  strict = FALSE,
  join = c("inner", "outer")
)

Arguments

Details

join = "outer" keeps the union and fills absent features with zero. A zero there means "not in this matrix", not "measured as zero".

Dense and sparse inputs may be mixed. Dense outer joins exceeding 5e7 elements use sparse storage to bound memory when features barely overlap.

Value

A matrix, dgCMatrix if the inputs were sparse.

Examples

## Not run: 
counts <- SeqoutCounts("GSE291735")
merged <- BindCounts(Matrices(counts, sample = wt$sample),
  labels = wt$stage, max_cells = 1200
)

## End(Not run)

Description

Queries against the cached table use local storage.

Usage

cache_table(table, con = .con())

CacheTable(table, con = .con())

Arguments

Value

The local table name (invisibly).

Description

Assembles entries from the dataset publication record. type = "all" includes reanalysis papers.

Usage

citations(accession, type = "original", con = .con())

Citations(accession, type = "original", con = .con())

Arguments

Details

A dataset with no linked paper returns character(0).

Reads REST. The dump lacks reanalysis papers and often publication dates.

Value

A character string of BibTeX entries, or character(0).

See Also

seqout_get(x)$pubs for the same papers as a tibble, and paper() to go the other way, from a paper to the datasets.

Examples

cat(Citations("GSE151530"))

# include reanalysis papers
cat(Citations("GSE168652", type = "all"))

writeLines(Citations("GSE151530"), file.path(tempdir(), "GSE151530.bib"))

Description

Removes ⁠*_local⁠ tables created by cache_table().

Usage

clear_cache(con = .con())

ClearCache(con = .con())

Arguments

Value

Number of tables removed (invisibly).

Description

Exact lookups against a fixed table; spellings the table lacks give NA.

Usage

country_code_to_name(code)

country_name_to_code(name)

CountryCodeToName(code)

CountryNameToCode(name)

Arguments

Value

A character vector the length of the input, NA where there is no match.

Examples

CountryCodeToName(c("IND", "USA", "XYZ"))
CountryNameToCode("India")

Description

Study counts per organism, assay category, archive, year and single-cell status for one country. Each facet holds its top 40 values.

Usage

country_facets(code, con = .con())

CountryFacets(code, con = .con())

Arguments

Value

A tibble with facet, value and studies columns.

See Also

country_projects().

Examples

f <- CountryFacets("IN")
f[f$facet == "organism", ]

Description

One row per study submitted from code. Filter values come from country_facets().

Usage

country_projects(
  code,
  organism = NULL,
  assay_l1 = NULL,
  source = NULL,
  has_fastq = NULL,
  has_sra = NULL,
  is_single_cell = NULL,
  q = NULL,
  sort = "year",
  order = "desc",
  limit = NULL,
  offset = 0,
  con = .con()
)

CountryProjects(
  code,
  organism = NULL,
  assay_l1 = NULL,
  source = NULL,
  has_fastq = NULL,
  has_sra = NULL,
  is_single_cell = NULL,
  q = NULL,
  sort = "year",
  order = "desc",
  limit = NULL,
  offset = 0,
  con = .con()
)

Arguments

Value

A study tibble, with a total attribute for the filtered count before limit cut it.

See Also

country_summary(), country_facets().

Examples

# single-cell studies submitted from India, newest first
CountryProjects("IN", is_single_cell = TRUE, limit = 100)

Description

code is an ISO-3166-1 alpha-2 code (e.g. "US", "IN"); matching is case-insensitive. A code that isn't two letters, or matches nothing, 404s.

Usage

country_summary(code, con = .con())

CountrySummary(code, con = .con())

Arguments

Value

A one-row tibble of totals.

See Also

country_facets() for the filter values, country_projects() for the studies themselves.

Examples

CountrySummary("US")

Description

The submitter's characteristics spread into one column each. Reuses the sample record seqout_counts() fetched.

Usage

counts_design(counts)

CountsDesign(counts)

Arguments

Value

A tibble, one row per sample of the study.

See Also

counts_samples() for the harmonised cohort fields.

Examples

## Not run: 
counts <- SeqoutListCounts("GSE135251")
col_data <- CountsDesign(counts)

## End(Not run)

Description

Keeps samples matching sample_search() with a readable unit.

Usage

counts_samples(counts, ..., min_cell_count = 1L)

CountsSamples(counts, ..., min_cell_count = 1L)

Arguments

Value

Matching samples, sorted by cells, with unit and format.

See Also

seqout_counts() for every unit, unfiltered.

Examples

## Not run: 
counts <- seqout_counts("GSE182159")
liver <- counts_samples(counts, tissue = "liver", min_cell_count = 1000)
m <- seqout_matrix(counts, sample = liver$unit[1])

## End(Not run)

Description

Matches q against the GARD and NORD catalogue names and their aliases, so a colloquial or older name finds the MONDO ids disease_projects() can then filter on.

Usage

disease_aliases(q, limit = 20, con = .con())

DiseaseAliases(q, limit = 20, con = .con())

Arguments

Value

A tibble with display_name, alias, mondo_ids and sources (list columns), with a total attribute for all matches and truncated for whether total exceeds limit.

See Also

disease_projects().

Examples

DiseaseAliases("marfan")

Description

Study counts per disease category/specialty/group (curated collections) or organism/assay/journal/country/year/status (a free-text term); see disease_summary() for the two modes.

Usage

disease_facets(collection, con = .con())

DiseaseFacets(collection, con = .con())

Arguments

Value

A tibble with facet, value and studies columns.

See Also

disease_projects().

Examples

DiseaseFacets("rare")

Description

One row per study, filtered and sorted; see disease_summary() for what distinguishes a curated collection from a free-text term. A curated row carries the catalogue match (catalogue_diseases, mondo_ids), per-scope sample counts, and stated/read-derived sex tallies; category/ specialty/group only apply to "rare", nord_type only to "nord". A term row carries the study's usual metadata plus n_samples_with_disease, the count of samples resolved to collection specifically (as opposed to n_samples, the study's total).

Usage

disease_projects(
  collection,
  category = NULL,
  specialty = NULL,
  group = NULL,
  nord_type = NULL,
  ancestry = NULL,
  inheritance = NULL,
  disease = NULL,
  assay_category = NULL,
  organism = NULL,
  assay_l1 = NULL,
  source = NULL,
  has_fastq = NULL,
  has_sra = NULL,
  is_single_cell = NULL,
  is_long_read = NULL,
  q = NULL,
  scope = "human_primary",
  sort = NULL,
  order = "desc",
  limit = NULL,
  offset = 0,
  con = .con()
)

DiseaseProjects(
  collection,
  category = NULL,
  specialty = NULL,
  group = NULL,
  nord_type = NULL,
  ancestry = NULL,
  inheritance = NULL,
  disease = NULL,
  assay_category = NULL,
  organism = NULL,
  assay_l1 = NULL,
  source = NULL,
  has_fastq = NULL,
  has_sra = NULL,
  is_single_cell = NULL,
  is_long_read = NULL,
  q = NULL,
  scope = "human_primary",
  sort = NULL,
  order = "desc",
  limit = NULL,
  offset = 0,
  con = .con()
)

Arguments

Details

has_fastq/has_sra are NA when the study is absent from the download-links table (unknown, not unavailable); filtering on them accepts only TRUE for that reason. scope (curated only) defaults to "human_primary" (human material that is neither an immortalized line nor a derived model); use "patient_derived_model", "cell_line", or "all" to see what that excludes.

Value

A study tibble, with a total attribute for the filtered count before limit cut it.

See Also

disease_summary(), disease_facets().

Examples

DiseaseProjects("rare", category = "Cancers", limit = 100)
DiseaseProjects("fatty liver disease", organism = "Homo sapiens", limit = 10)

Description

collection is either a curated catalogue name ("rare" for NIH GARD, "nord" for NORD) or any other string, resolved as free text against the full MONDO ontology and expanded to its descendants. The two modes return different columns: a curated summary carries per-scope study counts and stated/read-derived sex tallies; a term summary carries term/resolution/matched_labels describing how the text resolved, plus matched_samples scoped to exactly the resolved MONDO ids (samples is every sample in a matching study, not just the ones that resolved).

Usage

disease_summary(collection, con = .con())

DiseaseSummary(collection, con = .con())

Arguments

Details

A free-text term with no MONDO match 404s.

Value

A one-row tibble. Columns depend on collection; see Details.

See Also

disease_facets() for the filter values, disease_projects() for the studies themselves.

Examples

DiseaseSummary("rare")
DiseaseSummary("fatty liver disease")

Description

Fetches submitter BAMs aligned to their chosen reference.

Usage

download_bams(accession, dest_dir, quiet = FALSE, con = .con())

DownloadBAMs(accession, dest_dir, quiet = FALSE, con = .con())

Arguments

Details

Requester-pays files are named with the command to fetch them.

Value

The paths of the downloaded files, invisibly.

See Also

download_runs() for the reads, and seqout_get(x)$bams to see what exists before fetching any of it.

Examples

## Not run: 
SeqoutGet("SRP071083")$bams # requester-pays inventory
DownloadBAMs("SRP071083", "SRP071083")

## End(Not run)

Description

Fetches .parquet files for the Parquet backend. Existing files are skipped. HEAD sizes are reported before transfer when available.

Usage

download_dump(
  con = .con(),
  dest_dir,
  tables = NULL,
  overwrite = FALSE,
  quiet = FALSE
)

DownloadDump(
  con = .con(),
  dest_dir,
  tables = NULL,
  overwrite = FALSE,
  quiet = FALSE
)

Arguments

Value

The paths of the downloaded files, invisibly.

See Also

seqout_connect() to read the result, and cache_table() to keep one table in local DuckDB storage.

Examples

## Not run: 
# two common tables
DownloadDump(dest_dir = "~/seqout-dump", tables = c("unified_metadata", "geo_series"))

con <- SeqoutConnect("parquet", data_dir = "~/seqout-dump")
Query("SELECT count(*) FROM unified_metadata", con = con)

## End(Not run)

Description

Runs download their own files; larger accessions download all resolved runs. Default mode order is fastq, full SRA, then SRA lite. Published checksums are verified.

Usage

download_runs(accession, dest_dir, mode = NULL, quiet = FALSE, con = .con())

DownloadRuns(accession, dest_dir, mode = NULL, quiet = FALSE, con = .con())

Arguments

Value

The paths of the downloaded files, invisibly.

Examples

## Not run: 
DownloadRuns("SRR12012336", "SRR12012336") # one run
DownloadRuns("SRP267291", "SRP267291") # every run of a study
DownloadRuns("SRP150719", "SRP150719", mode = "fastq") # insist on ENA fastq

## End(Not run)

Description

Downloads processed files: matrices, annotations and archives. Samples stay scoped to their own files.

Usage

download_supplementary(accession, dest_dir, quiet = FALSE, con = .con())

DownloadSupplementary(accession, dest_dir, quiet = FALSE, con = .con())

Arguments

Value

The paths of the downloaded files, invisibly.

Examples

## Not run: 
DownloadSupplementary("GSE168652", "GSE168652") # series files and every sample's
DownloadSupplementary("GSM8433846", "GSM8433846") # that one sample's
DownloadSupplementary("E-MTAB-11467", "E-MTAB-11467") # processed archives and raw

## End(Not run)

Description

Names the role from the filename, before download.

Usage

file_role(name)

FileRole(name)

Arguments

Value

One of "mtx", "barcodes", "features", "h5", "h5ad", "rds", "rcc", "table", "tar", "metadata" or "skip".

Examples

FileRole("GSM123_matrix.mtx.gz")
FileRole("GSM123_barcodes.tsv.gz")
FileRole("GSE1_cell_metadata.csv.gz")
FileRole("GSM1_fragments.tsv.gz")
FileRole("GSM4567420_01_NAS5.RCC.gz")

Description

Enumerates values accepted by seqout_search(), with record counts.

Usage

list_library_strategies(con = .con())

list_instrument_models(con = .con())

list_journals(con = .con(), limit = 500)

list_centers(con = .con(), limit = 500)

list_organisms(con = .con(), common_names = FALSE)

list_assays(con = .con(), country = NULL)

list_platforms(con = .con())

ListAssays(con = .con(), country = NULL)

ListCenters(con = .con(), limit = 500)

ListInstrumentModels(con = .con())

ListJournals(con = .con(), limit = 500)

ListLibraryStrategies(con = .con())

ListOrganisms(con = .con(), common_names = FALSE)

ListPlatforms(con = .con())

Arguments

Details

list_assays() returns both assay levels in a level column. These functions read REST; query() can count from the dump.

Value

A tibble of values and counts.

See Also

seqout_search().

Examples

ListLibraryStrategies()

# instrument_model values
ListInstrumentModels()

# connection-first call
SeqoutConnect("api") |> ListJournals(limit = 20)

## Not run: 
# pass a value into search
top <- ListLibraryStrategies()$value[1]
SeqoutSearch("liver", library_strategy = top, limit = 10)

## End(Not run)

Description

CellRanger directories win; otherwise strip compression and role tokens. Matching stems place matrix.mtx, barcodes.tsv, and features.tsv together.

Usage

group_key(name)

GroupKey(name)

Arguments

Value

The grouping key.

Examples

GroupKey("GSM123_pbmc_matrix.mtx.gz")
GroupKey("GSM123/filtered_feature_bc_matrix/matrix.mtx.gz")

Description

Find the GEO series a sample belongs to

Usage

gsm_series(gsm, con = .con())

GSMSeries(gsm, con = .con())

Arguments

Value

The GSE accession, or NA_character_.

Description

Filtered output ranks before raw.

Usage

is_filtered(name)

IsFiltered(name)

Arguments

Value

TRUE when the name marks filtered output.

Examples

IsFiltered("GSM1_filtered_feature_bc_matrix.h5")
IsFiltered("GSM1_raw_feature_bc_matrix.h5")

Description

Resolves a sequencing study to its GEO or ArrayExpress series.

Usage

linked_geo(accession, con = .con())

LinkedGEO(accession, con = .con())

Arguments

Value

The linked series accession, or NA_character_.

Description

GEO and ArrayExpress runs live under linked sequencing studies. GEA resolves through BioProject.

Usage

linked_study(accession, con = .con())

LinkedStudy(accession, con = .con())

Arguments

Value

The linked study accession, or NA_character_.

Description

Study counts per technology, platform, instrument model, library strategy, organism, archive, chemistry and year.

Usage

longread_facets(con = .con())

LongreadFacets(con = .con())

Arguments

Value

A tibble with facet, value and studies columns.

See Also

longread_projects().

Examples

f <- LongreadFacets()
f[f$facet == "technology", ]

Description

One row per study; a study mirrored in more than one archive counts once. long_read_only = TRUE had no other platform; FALSE is hybrid; NA has no experiment rows to judge by (GEO-only entries). Filter values come from longread_facets().

Usage

longread_projects(
  technology = NULL,
  platform = NULL,
  instrument_model = NULL,
  library_strategy = NULL,
  organism = NULL,
  archive = NULL,
  chemistry = NULL,
  assay_l1 = NULL,
  year = NULL,
  has_fastq = NULL,
  has_sra = NULL,
  long_read_only = NULL,
  has_exact_chemistry = NULL,
  single_cell = NULL,
  sort = "n_experiments",
  order = "desc",
  limit = NULL,
  offset = 0,
  con = .con()
)

LongreadProjects(
  technology = NULL,
  platform = NULL,
  instrument_model = NULL,
  library_strategy = NULL,
  organism = NULL,
  archive = NULL,
  chemistry = NULL,
  assay_l1 = NULL,
  year = NULL,
  has_fastq = NULL,
  has_sra = NULL,
  long_read_only = NULL,
  has_exact_chemistry = NULL,
  single_cell = NULL,
  sort = "n_experiments",
  order = "desc",
  limit = NULL,
  offset = 0,
  con = .con()
)

Arguments

Details

single_cell keeps only studies also flagged single-cell (is_single_cell). The server has no such parameter, so this filters locally.

Value

A study tibble, with a total attribute for the filtered count before limit cut it.

See Also

longread_summary(), longread_facets(), project_longread_chemistry() for one study's per-run detail.

Examples

# long-read-only datasets
LongreadProjects(long_read_only = TRUE, limit = 10)

# long-read AND single-cell
LongreadProjects(single_cell = TRUE, limit = 10)

# Oxford Nanopore human studies, newest first
LongreadProjects(
  technology = "Oxford Nanopore", organism = "Homo sapiens",
  sort = "first_published", limit = 10
)

Description

Totals for studies with at least one PacBio or Oxford Nanopore experiment, across GEO, SRA, ENA, DDBJ DRA and GSA. A study mirrored in more than one archive counts once. studies_hybrid also sequenced on a short-read platform; studies_long_read_only did not; the remainder have no experiment rows to judge by (GEO-only entries).

Usage

longread_summary(con = .con())

LongreadSummary(con = .con())

Arguments

Value

A one-row tibble of totals.

See Also

longread_facets() for the filter values, longread_projects() for the studies themselves.

Examples

LongreadSummary()

Description

Adds ⁠<column>_ontology_id⁠ columns with comma-separated CURIEs. Missing or empty labels become NA.

Usage

map_to_ontology(
  x,
  columns,
  ontology = NULL,
  use_synonyms = FALSE,
  max_hops = 1,
  con = .con()
)

MapToOntology(
  x,
  columns,
  ontology = NULL,
  use_synonyms = FALSE,
  max_hops = 1,
  con = .con()
)

Arguments

Details

Only the identifiers of the label itself are read. Under use_synonyms, synonym IDs can be narrower; own IDs win.

Reads REST. The ontology graph is absent from the dump.

Value

x with one new character column for each column that you named.

See Also

ontology() for what the graph holds about one term.

Examples

meta <- data.frame(celltype = c("T cell", "hepatocyte", "HeLa"))
MapToOntology(meta, "celltype")

# Cell Ontology IDs
MapToOntology(meta, "celltype", ontology = "CL")

# allow synonym IDs
MapToOntology(meta, "celltype", use_synonyms = TRUE)

Description

Broken units are skipped with a warning.

Usage

matrices(counts, sample = NULL)

Matrices(counts, sample = NULL)

Arguments

Value

A named list of seqout_matrix objects, keyed by unit label.

Description

Flattens the microbes list column from sample_search() into one row per (sample, organism, run).

Usage

microbe_detections(
  x,
  validated_only = FALSE,
  columns = c("organism", "class", "kingdom", "breadth_frac", "kmer_mass")
)

MicrobeDetections(
  x,
  validated_only = FALSE,
  columns = c("organism", "class", "kingdom", "breadth_frac", "kmer_mass")
)

Arguments

Value

A tibble of sample plus the requested detection fields, empty when nothing was detected.

See Also

sample_search() for the samples, sample_microbes() for every detection in one sample.

Examples

hpv <- SampleSearch(
  disease_ontology_id = "MONDO:0002974", microbe = "HPV", limit = 10
)
d <- MicrobeDetections(hpv)
sort(table(d$organism), decreasing = TRUE)

# only the detections that pass the gates
MicrobeDetections(hpv, validated_only = TRUE)

Description

Looks up one term in the graph used for search expansion. Returns source IDs, synonyms and direct children.

Usage

ontology(term, max_hops = 2, children = TRUE, con = .con())

Ontology(term, max_hops = 2, children = TRUE, con = .con())

Arguments

Details

max_hops bounds the walk over the synonym links only. Children are always the direct children of the resulting synonym cluster, at any max_hops.

A term absent from the graph returns an empty tibble.

Reads REST. The ontology graph is absent from the dump.

Value

A tibble with one row for the term, one for each synonym and one for each child. Columns: relation, name, xrefs (a list column of source CURIEs) and has_children. The synonym_total, children_truncated and max_hops attributes report what the server capped.

See Also

seqout_search(), which expands a plain query through this graph.

Examples

Ontology("liver")

# The identifiers alone, without the hierarchy query
Ontology("breast cancer", children = FALSE)$xrefs

# Terms one level down that expand further
onto <- Ontology("liver")
onto[onto$relation == "child" & onto$has_children, "name"]

Description

Find projects linked to a publication

Usage

paper(id = NULL, pmid = NULL, doi = NULL, con = .con())

Paper(id = NULL, pmid = NULL, doi = NULL, con = .con())

Arguments

Value

A tibble of linked projects, empty when the publication is unknown.

Examples

Paper("34764296")
Paper(doi = "10.1038/s41467-021-26864-x")

Description

Study counts per perturbation type, confidence, data availability, genetic tool, perturbation method, compound, organism, tissue, readout assay and year. Each facet holds its top 100 values.

Usage

perturbation_facets(con = .con())

PerturbationFacets(con = .con())

Arguments

Value

A tibble with facet, value and studies columns.

See Also

perturbation_projects().

Examples

f <- PerturbationFacets()
f[f$facet == "perturbation_method", ]

Description

One row per single-cell study with genetic (CRISPR knockout, CRISPRi, CRISPRa, RNAi, ORF, base or prime editing), chemical (drug or biologic) or other-stimulus perturbation evidence, with the readout assay and whether a counted matrix, raw FASTQ, both or neither is available.

Usage

perturbation_projects(
  perturbation_type = NULL,
  confidence = NULL,
  min_confidence = NULL,
  data_availability = NULL,
  genetic_subtype = NULL,
  perturbation_method = NULL,
  compound = NULL,
  readout_assay = NULL,
  cell_line = NULL,
  organism = NULL,
  tissue = NULL,
  year = NULL,
  has_matrix = NULL,
  has_fastq = NULL,
  has_control_arm = NULL,
  is_pooled = NULL,
  is_long_read = NULL,
  q = NULL,
  sort = "confidence",
  order = "desc",
  limit = NULL,
  offset = 0,
  con = .con()
)

PerturbationProjects(
  perturbation_type = NULL,
  confidence = NULL,
  min_confidence = NULL,
  data_availability = NULL,
  genetic_subtype = NULL,
  perturbation_method = NULL,
  compound = NULL,
  readout_assay = NULL,
  cell_line = NULL,
  organism = NULL,
  tissue = NULL,
  year = NULL,
  has_matrix = NULL,
  has_fastq = NULL,
  has_control_arm = NULL,
  is_pooled = NULL,
  is_long_read = NULL,
  q = NULL,
  sort = "confidence",
  order = "desc",
  limit = NULL,
  offset = 0,
  con = .con()
)

Arguments

Details

Detection is rule-based, from named methods in the study's own text (Perturb-seq, CROP-seq, ECCITE-seq, Mosaic-seq, TAP-seq, Spear-ATAC, sci-Plex, MIX-Seq and others), guide-library and pooled-screen wording, a control-arm design test on sample treatment values, and a ChEBI compound lexicon.

confidence is "high", "medium" or "low". High needs a named method, a guide library with screen text, or several compound treatments beside a control arm. Medium is corroborated by a second signal. Low is a single weak signal and includes false positives, so start from min_confidence = "medium" when you want studies to rely on. evidence lists the signals behind each row. perturbation_type is "genetic", "chemical", "both" or "other" (a cytokine or stimulation design with a control). Germline knockout and transgenic models are not evidence, and doxycycline or tamoxifen used as Cre or Tet inducers never count as the perturbation.

compounds are ChEBI-matched names from sample treatment values. They are empty when the drug arms live only in supplementary files or cell barcodes, as in sci-Plex, so a named method with no compounds is expected. data_availability is "both", "matrix_only", "fastq_only" or "neither", from has_matrix and has_fastq. Filter values come from perturbation_facets().

Value

A study tibble, with a total attribute for the filtered count before limit cut it.

See Also

perturbation_summary(), perturbation_facets(), seqout_counts() to read the matrices of a study with has_matrix.

Examples

# studies to rely on that have both a matrix and FASTQ
PerturbationProjects(
  min_confidence = "medium", data_availability = "both", limit = 100
)

# CRISPRi screens in human
PerturbationProjects(
  perturbation_type = "genetic", genetic_subtype = "CRISPRi",
  organism = "Homo sapiens", min_confidence = "medium", limit = 10
)

# drug studies that name trametinib
PerturbationProjects(compound = "trametinib", has_matrix = TRUE, limit = 10)

Description

Totals for single-cell studies with genetic or chemical perturbation evidence, the same population perturbation_projects() lists. studies_high_medium counts the studies whose evidence is strong enough to rely on; studies also counts weak single-signal calls.

Usage

perturbation_summary(con = .con())

PerturbationSummary(con = .con())

Arguments

Value

A one-row tibble of totals.

See Also

perturbation_facets() for the filter values, perturbation_projects() for the studies themselves.

Examples

PerturbationSummary()

Description

Every PacBio/Oxford Nanopore run in the study, independent of single-cell status and Pentimento linkage. chemistry_confidence is exact when read from the submitted BAM's own header, declared (ONT only) when parsed from the depositor's own protocol text, bucket when inferred from instrument_model alone, unknown otherwise.

Usage

project_longread_chemistry(accession, con = .con())

ProjectLongreadChemistry(accession, con = .con())

Arguments

Value

A run tibble, empty when the study has no long-read runs.

See Also

project_single_cell(), whose longread_chemistry attribute answers the same question restricted to the study's first page of samples.

Examples

ProjectLongreadChemistry("GSE297547")

Description

One row per sample with matrix dimensions and read-derived calls.

Usage

project_single_cell(accession, limit = NULL, offset = 0, con = .con())

ProjectSingleCell(accession, limit = NULL, offset = 0, con = .con())

Arguments

Details

cells counts matrix columns; for unfiltered 10x matrices these are barcodes, so the number is an upper bound and sums overcount.

⁠has_*_reads = NA⁠ means unscreened; FALSE means screened with no hit. flags preserves the same distinction.

Value

A sample tibble, empty when the study is unprocessed. The study and n_samples_total attributes carry the study row and its total. The longread_chemistry attribute carries every PacBio/Oxford Nanopore run in the study (empty, not NULL, when there are none); the server only computes it when offset = 0.

See Also

sample_microbes() for the per-organism detections behind the viral and bacterial flags, project_longread_chemistry() to read the same run list without an offset argument.

Examples

sc <- ProjectSingleCell("GSE168652")
sc[, c("sample_accession", "cells", "genes")]
attr(sc, "study")$study_cells
attr(sc, "longread_chemistry")

Description

kind describes the evidence behind the call:

matrix_and_reads

counts and reads, both confirmed

matrix_reads_unscanned

counts plus linked FASTQ no read-derived call has confirmed yet

matrix_only

counts, no linked reads

reads_only

no parsed matrix

Usage

project_single_cell_status(accession, con = .con())

ProjectSingleCellStatus(accession, con = .con())

Arguments

Value

A one-row tibble, or an empty one when the accession is unknown to both the unified catalogue and the Pentimento.

See Also

project_single_cell_summary() for the study-level rollup, project_single_cell() for the per-sample counts.

Examples

st <- ProjectSingleCellStatus("GSE168652")
st$single_cell_status
st$n_independent_evidence

Description

Study-level single-cell summary: cells, reads scanned, microbes

Usage

project_single_cell_summary(accession, con = .con())

ProjectSingleCellSummary(accession, con = .con())

Arguments

Value

A one-row tibble, empty when the study is not in the Pentimento.

See Also

project_single_cell_status() when you need an answer even for a study absent from the Pentimento.

Examples

s <- ProjectSingleCellSummary("GSE168652")
s$study_cells
s$microbe_measured

Description

The DuckDB database includes remote views and cached local tables.

Usage

query(sql, params = NULL, con = .con())

Query(sql, params = NULL, con = .con())

Arguments

Value

A tibble with query results.

Examples

## Not run: 
con <- SeqoutConnect("parquet")
Query("SELECT accession, title FROM geo_series LIMIT 10", con = con)

SeqoutDefault(con) # or make it the session default
Query("
  SELECT dominant_scientific_name AS organism, count(*) AS n
  FROM unified_metadata
  WHERE dominant_scientific_name IS NOT NULL
  GROUP BY organism
  ORDER BY n DESC
  LIMIT 20
")

## End(Not run)

Description

Scores markers per cell, averages by cluster, and labels by the top set. Marker sets with no genes in x are dropped with a warning.

Usage

quick_annotation(x, clusters, markers, normalize = "auto")

QuickAnnotation(x, clusters, markers, normalize = "auto")

Arguments

Details

Scores are unscaled; housekeeping-heavy sets can win.

Value

A character vector of labels, named by cluster, with the cluster by marker-set score matrix attached as attribute scores.

Examples

## Not run: 
markers <- list(
  neuron = c("Neurod2", "Tbr1"),
  microglia = c("Cx3cr1", "C1qa")
)
expr <- Seurat::GetAssayData(obj, assay = "RNA", slot = "data") # Seurat v4
expr <- SeuratObject::LayerData(Seurat::JoinLayers(obj), layer = "data") # v5

labels <- QuickAnnotation(expr, obj$seurat_clusters, markers)
obj$celltype <- unname(labels[as.character(obj$seurat_clusters)])

## End(Not run)

Description

Views are lazy; catalog reads need them registered first.

Usage

register_tables(con = .con(), tables = NULL, progress = interactive())

RegisterTables(con = .con(), tables = NULL, progress = interactive())

Arguments

Value

The names of every registered view, invisibly.

Examples

## Not run: 
con <- SeqoutConnect("parquet")
RegisterTables(con, "unified_metadata")
dplyr::tbl(con$db, "unified_metadata")

# all views
SeqoutConnect("parquet") |> RegisterTables()

## End(Not run)

Description

Resolves a run, experiment or sample to its study or series.

Usage

resolve_study(accession, con = .con())

ResolveStudy(accession, con = .con())

Arguments

Value

The study accession, or NA_character_ when nothing links back.

Examples

resolve_study("SRR13927092")

Description

One row per file: a paired run gives two. Uses the same choice of copy as download_runs(), so this is what that call would fetch.

Usage

run_files(x, mode = NULL, con = .con())

RunFiles(x, mode = NULL, con = .con())

Arguments

Value

A tibble with run, mode, url, name (the file name on disk), bytes and md5; bytes/md5 are NA where the archive publishes none. Runs with no downloadable copy are left out.

Examples

files <- RunFiles("SRP148597", mode = "fastq")
sum(files$bytes, na.rm = TRUE) / 1e9 # GB to fetch
"SRR7191948" |> RunFiles()

Description

Returns one row per organism, ordered by k-mer mass.

Usage

sample_microbes(
  accession,
  kind = c("all", "viral", "bacterial", "both"),
  min_breadth = NULL,
  min_kmer_mass = NULL,
  validated_only = FALSE,
  include_background = FALSE,
  detail = FALSE,
  limit = 500,
  con = .con()
)

SampleMicrobes(
  accession,
  kind = c("all", "viral", "bacterial", "both"),
  min_breadth = NULL,
  min_kmer_mass = NULL,
  validated_only = FALSE,
  include_background = FALSE,
  detail = FALSE,
  limit = 500,
  con = .con()
)

Arguments

Details

Panel-unitig alignments lack raw-read, UMI and replicate confirmation. measurable = FALSE or NA means unscreened and triggers a warning.

kmer_mass weights k-mers within a run, so it carries no absolute scale. n_unitigs counts mapped unitigs.

Spike-in, negative-control and background classes are excluded from totals.

Value

An organism tibble, or a run-by-organism tibble when detail = TRUE. measurable, n_runs, totals, by_kingdom, control_kingdoms and the unpicked view are carried as attributes.

See Also

project_single_cell() for the study-wide flags.

Examples

m <- SampleMicrobes("GSM5155196", kind = "viral")
m[, c("organism", "n_unitigs", "max_breadth_frac")]
attr(m, "measurable")
attr(m, "detections")

Description

Searches harmonised sample metadata. Unharmonised samples are absent.

Usage

sample_search(
  filters = NULL,
  ...,
  single_cell = NULL,
  include_descendants = TRUE,
  sort = "sample",
  order = "asc",
  limit = NULL,
  con = .con()
)

SampleSearch(
  filters = NULL,
  ...,
  single_cell = NULL,
  include_descendants = TRUE,
  sort = "sample",
  order = "asc",
  limit = NULL,
  con = .con()
)

Arguments

Details

Substring

tissue, disease, cell_type, assay, assay_category, phenotype, treatment, development_stage, sample_type, genetic_modification, strain, cell_line, ethnicity, tissue_primary_site. "liver" matches "liver, left lobe".

Exact

organism, sex, taxid, study_accession. Exact matching keeps "male" from matching "female".

Harmonised ontology ID

disease_ontology_id, tissue_ontology_id, cell_type_ontology_id, assay_ontology_id, development_stage_ontology_id. Give a CURIE such as "MONDO:0005061". See include_descendants.

Range

age_min_years, age_max_years, min_cell_count, max_cell_count, min_gene_count, max_gene_count. An age filter excludes a sample without a recorded age, so age_min_years = 0 means "has a recorded age".

Matrix

has_matrix keeps samples with a parsed cell-by-gene matrix.

Read-derived

single_cell_only (but see single_cell), has_viral_reads, has_bacterial_reads, hpv_type, microbe, microbe_class, microbe_min_breadth, microbe_min_kmer_mass, microbe_validated_only.

A ⁠microbe*⁠ filter keeps samples with matching detections and attaches them.

Requires REST. The harmonised sample table is absent from the Parquet dump.

Value

A sample tibble with total and filters attributes.

See Also

sample_microbes() for every detection in one sample, seqout_search() when the answer should be projects.

Examples

# single-cell samples by assay text
SampleSearch(tissue = "liver", disease = "steato", single_cell = "assay", limit = 10)

# female human liver samples over 50
SampleSearch(
  organism = "Homo sapiens", sex = "female",
  tissue = "liver", age_min_years = 50, limit = 10
)

# HPV references use HPV16, HPV18, and related names
# "papillomavirus" matches nothing
hpv <- SampleSearch(
  tissue = "cervix", microbe = "HPV",
  sort = "cell_count", order = "desc", limit = 10
)
attr(hpv, "total")

# filters built in code
list(tissue = "cervix", microbe = "HPV") |> SampleSearch(limit = 10)

# ontology term and subtypes
SampleSearch(disease_ontology_id = "MONDO:0005061", limit = 10)

# exact term
SampleSearch(
  disease_ontology_id = "MONDO:0005061",
  include_descendants = FALSE, limit = 10
)

Description

Groups the full seqout_search() match set by facet.

Usage

search_facets(
  query,
  ...,
  structured = FALSE,
  exclude_ontology = NULL,
  con = .con()
)

SearchFacets(
  query,
  ...,
  structured = FALSE,
  exclude_ontology = NULL,
  con = .con()
)

Arguments

Details

Values order by score, the summed match rank. score is 0 without a query.

Value

A tibble of facet, value, count and score, with total and max_rank attributes.

See Also

seqout_search() for the results themselves, search_suggest() when a query returns nothing.

Examples

f <- SearchFacets("liver cancer")
attr(f, "total")

# organism facet
f[f$facet == "organism", ]

# count within a narrower set
SearchFacets("liver cancer", organism = "Homo sapiens")

Description

Suggest spelling corrections for a query

Usage

search_suggest(query, con = .con())

SearchSuggest(query, con = .con())

Arguments

Value

A tibble of suggestions, empty when the query needs no correction.

See Also

seqout_search(), and search_facets() for a query that matched too much.

Examples

SearchSuggest("livre cancr")

# already spelled correctly
SearchSuggest("liver cancer")

Description

Parquet closes its DuckDB handle; REST has no handle.

Usage

seqout_close(con = .con())

SeqoutClose(con = .con())

Arguments

Value

NULL, invisibly. Called for its side effect.

Description

Two backends are supported:

Usage

seqout_connect(
  backend = c("api", "parquet"),
  base_url = Sys.getenv("SEQOUT_BASE_URL", "https://seqout.org"),
  data_dir = NULL,
  read_only = FALSE,
  eager = FALSE,
  quiet = FALSE
)

SeqoutConnect(
  backend = c("api", "parquet"),
  base_url = Sys.getenv("SEQOUT_BASE_URL", "https://seqout.org"),
  data_dir = NULL,
  read_only = FALSE,
  eager = FALSE,
  quiet = FALSE
)

Arguments

Details

• "api", the default, reads https://seqout.org with HTTP.

• "parquet" reads the Parquet dump with DuckDB through httpfs. It answers SQL and can read a local dump. Requires the duckdb package.

Select Parquet explicitly; one lookup can become a dump scan.

Value

A seqout_connection object.

Examples

## Not run: 
# REST API default
project("GSE297547")

# Parquet for SQL and local dumps
con <- SeqoutConnect("parquet")
Query("SELECT * FROM geo_series LIMIT 5", con = con)
SeqoutClose(con)

# local dump
con <- SeqoutConnect("parquet", data_dir = "~/seqout-dump")
Query("SELECT count(*) FROM geo_series", con = con)
SeqoutClose(con)

## End(Not run)

Description

Uses genevintage::correct_genenames() to resolve date-like gene names against the chosen species and annotation. Tokens with multiple candidates or no candidate are retained and reported. This cannot reconstruct every Excel error or recover information lost through rounding.

Usage

seqout_correct_names(
  x,
  ids = NULL,
  species = NULL,
  release = NULL,
  assembly = NULL,
  source = NULL,
  mapping = NULL,
  quiet = FALSE
)

SeqoutCorrectNames(
  x,
  ids = NULL,
  species = NULL,
  release = NULL,
  assembly = NULL,
  source = NULL,
  mapping = NULL,
  quiet = FALSE
)

Arguments

Details

Supply mapping, or both species and release; genevintage refuses damaged input otherwise, since symbols alone cannot establish species or annotation vintage. Choosing a release explicitly also distinguishes historical names such as SEPT9 from SEPTIN9. Only call this on gene labels: real dates and numeric identifiers can look like Excel damage. Neither this function nor the matrix reader repairs arbitrary spreadsheet cells.

Value

See seqout_gene_names(). Status is "corrected", "ambiguous", "unresolved", or "unchanged". The audit preserves damaged input tokens.

See Also

seqout_gene_names()

Examples

x <- matrix(1:4,
  nrow = 2,
  dimnames = list(c("9-Sep", "TP53"), c("s1", "s2"))
)
mapping <- data.frame(
  id = c("ENSG00000184640", "ENSG00000141510"),
  name = c("SEPTIN9", "TP53"), chr = "17", biotype = "protein_coding"
)
SeqoutCorrectNames(x, mapping = mapping)

Description

Groups supplementary files into readable units without downloading them.

Usage

seqout_counts(
  accession,
  assay = "rna",
  feature_type = NULL,
  cache_dir = NULL,
  con = .con()
)

SeqoutCounts(
  accession,
  assay = "rna",
  feature_type = NULL,
  cache_dir = NULL,
  con = .con()
)

SeqoutListCounts(
  accession,
  assay = "rna",
  feature_type = NULL,
  cache_dir = NULL,
  con = .con()
)

seqout_list_counts(
  accession,
  assay = "rna",
  feature_type = NULL,
  cache_dir = NULL,
  con = .con()
)

Arguments

Details

A unit is a 10x triplet, CellRanger .h5, .h5ad, .rds, tar, or table.

has_metadata means embedded metadata or a named sidecar was found.

preferred ranks requested assay, filtered output, then format.

10x feature class follows assay; feature_type overrides it.

Value

A tibble, one row per readable unit.

Examples

## Not run: 
counts <- SeqoutListCounts("GSE297547")
counts[, c("unit", "sample", "format", "preferred")]
m <- SeqoutMatrix(counts, sample = "GSM8994520")

"GSE297547" |>
  SeqoutListCounts() |>
  SeqoutMatrix(sample = "GSM8994520")

## End(Not run)

Description

Download a unit's files without parsing them

Usage

seqout_counts_files(counts, sample = NULL)

SeqoutCountsFiles(counts, sample = NULL)

Arguments

Value

The paths of the downloaded files.

Description

Set the process-wide default connection

Usage

seqout_default(con)

SeqoutDefault(con)

Arguments

Value

The previous default, invisibly.

Examples

## Not run: 
SeqoutDefault(SeqoutConnect("parquet"))
project("GSE297547") # now reads Parquet
SeqoutDefault(NULL) # back to REST

## End(Not run)

Description

A fragments file holds one bgzipped BED-like row per ATAC fragment, with columns chrom, start, end, barcode and read_support.

Usage

seqout_fragments(counts, sample = NULL, download = FALSE)

SeqoutFragments(counts, sample = NULL, download = FALSE)

Arguments

Details

Files download only when download = TRUE.

GEO rarely holds the .tbi index Signac wants. Build one after downloading with Rsamtools::indexTabix(path, format = "bed").

That works on a CellRanger fragments file, which is bgzipped and sorted by coordinate. Some deposits are neither: GSE184462 ships plain-gzipped files sorted by barcode, and tabix rejects them with "Chromosome blocks not continuous". Sort and recompress first: ⁠sort -k1,1 -k2,2n⁠, then Rsamtools::bgzip(), then indexTabix().

Value

A tibble of sample, file, url, indexed and path. path is NA until the file is downloaded.

See Also

seqout_counts() for the matrices, seqout_counts_files() for a unit's own files.

Examples

## Not run: 
counts <- SeqoutListCounts("GSE156478")
SeqoutFragments(counts)

frags <- SeqoutFragments(counts, sample = "GSM5065524", download = TRUE)
Rsamtools::indexTabix(frags$path[1], format = "bed")

## End(Not run)

Description

Uses genevintage::geneid2name() to detect species and rank compatible annotation releases, or genevintage::gene_names() with a supplied mapping. Detection is an estimate, especially for small or filtered gene sets. Pass the known annotation explicitly when available. Unmapped IDs are retained.

Usage

seqout_gene_names(
  x,
  ids = NULL,
  species = NULL,
  release = NULL,
  assembly = NULL,
  source = NULL,
  mapping = NULL,
  quiet = FALSE
)

SeqoutGeneNames(
  x,
  ids = NULL,
  species = NULL,
  release = NULL,
  assembly = NULL,
  source = NULL,
  mapping = NULL,
  quiet = FALSE
)

Arguments

Value

The same kind of object with renamed rows. Counts, row order and observation metadata are preserved. For a seqout_matrix, var row names are updated together with X; original annotation columns are preserved. An audit is appended to x$gene_annotation (or the gene_annotation attribute of a plain matrix). Each entry records operation, annotation and a features table with the original row, input token, resolved name, unique output row name and status. Colliding row names get .1, .2, etc.; the unsuffixed gene name remains in the audit. Rows are never aggregated.

See Also

seqout_correct_names(), genevintage::detect_release()

Examples

x <- matrix(1:4, nrow = 2, dimnames = list(
  c("ENSG00000141510.16", "ENSG00000012048.23"), c("s1", "s2")
))
mapping <- data.frame(
  id = c("ENSG00000141510", "ENSG00000012048"),
  name = c("TP53", "BRCA1")
)
SeqoutGeneNames(x, mapping = mapping)

Description

Reads the audit seqout_gene_names() or seqout_correct_names() appended, wherever it lives: x$gene_annotation for a seqout_matrix, the gene_annotation attribute for a plain matrix.

Usage

seqout_gene_report(x, which = NULL)

SeqoutGeneReport(x, which = NULL)

Arguments

Value

The audit's features table: original, input, name, row_name and status, one row per matrix row. The call's operation and annotation are attached as attributes of the same names.

Examples

x <- matrix(1:4, nrow = 2, dimnames = list(
  c("ENSG00000141510.16", "ENSG00000012048.23"), c("s1", "s2")
))
mapping <- data.frame(
  id = c("ENSG00000141510", "ENSG00000012048"),
  name = c("TP53", "BRCA1")
)
x |>
  SeqoutGeneNames(mapping = mapping) |>
  SeqoutGeneReport()

Description

Lazy fields fetch on first access and then cache their result.

Usage

seqout_get(accession, con = .con())

SeqoutGet(accession, con = .con())

Arguments

Details

The fields are:

meta

the project record: title, summary, design, dates, files

samples

one record per sample

experiments

one record per library preparation

runs

one record per sequencing run, with its file URLs

supplementary

one row per supplementary file, series and sample alike; sample is NA on the ones the series carries itself

bams

the alignment files a submitter sent, where there are any

pubs

the publications linked to the dataset

links

the same data in other archives

enriched

the harmonised data for the samples: the submitter's free text read into one vocabulary, with ontology IDs

detail

the record for a sample or run accession

Value

A seqout_dataset object with lazy fields.

Examples

d <- SeqoutGet("GSE168652")
d$meta$title
nrow(d$samples)
nrow(d$runs)

Description

Read one unit as a counts matrix

Usage

seqout_matrix(counts, sample = NULL)

SeqoutMatrix(counts, sample = NULL)

Arguments

Value

A seqout_matrix: a list with X (features by observations), obs, var, kind, fmt and evidence.

Description

Asks the REST root once per session with a short timeout, so scripts, examples and vignettes can skip cleanly while the service is down instead of failing on their first query.

Usage

seqout_online(con = .con())

SeqoutOnline(con = .con())

Arguments

Value

TRUE when the API answered, FALSE otherwise.

Examples

SeqoutOnline()

Description

X becomes counts; obs becomes colData; var becomes rowData. A GSE or GSM accession is resolved through seqout_counts(). Requires SingleCellExperiment.

Usage

seqout_sce(
  x,
  assay_name = "counts",
  sample = NULL,
  max_cells = NULL,
  multimodal = TRUE,
  sample_metadata = FALSE,
  ...
)

Seqout2SCE(
  x,
  assay_name = "counts",
  sample = NULL,
  max_cells = NULL,
  multimodal = TRUE,
  sample_metadata = FALSE,
  ...
)

Arguments

Value

A SingleCellExperiment object.

See Also

seqout_seurat(), seqout_matrix() and matrices().

Examples

## Not run: 
sce <- Seqout2SCE("GSM8994520")

sce <- Seqout2SCE("GSE297547", sample = "GSM8994520")
counts <- SeqoutListCounts("GSE297547")
sce <- Seqout2SCE(counts, sample = "GSM8994520")

## End(Not run)

Description

Searches GEO, SRA, ArrayExpress, ENA, GSA, DRA and GEA together.

Usage

seqout_search(
  query = NULL,
  ...,
  sortby = NULL,
  order = "desc",
  limit = NULL,
  structured = FALSE,
  expand = TRUE,
  exclude_ontology = NULL,
  con = .con()
)

SeqoutSearch(
  query = NULL,
  ...,
  sortby = NULL,
  order = "desc",
  limit = NULL,
  structured = FALSE,
  expand = TRUE,
  exclude_ontology = NULL,
  con = .con()
)

Arguments

Value

A tibble of results, with a took_ms attribute.

See Also

sample_search() when the answer should be samples, and the Search article for the filters and the costs.

Examples

SeqoutSearch("liver cancer scRNA", limit = 10)


# filters combine freely
SeqoutSearch("liver cancer", organism = "Homo sapiens", country = "Japan", limit = 10)

# exact words
SeqoutSearch("spinal muscular atrophy", expand = FALSE, limit = 10)


## Not run: 
# boolean syntax triggers structured parsing
SeqoutSearch('("aging" OR "aged") (gut OR colon) immun*', limit = 10)

# force structured parsing
SeqoutSearch("liver cancer", structured = TRUE, limit = 10)

# exclude ontology sources
SeqoutSearch("spinal muscular atrophy", exclude_ontology = c("MeSH", "CVCL"), limit = 10)

# rank or count a broad match set
SeqoutSearch("liver cancer scRNA", db = "geo", sortby = "citations", limit = 10)
SeqoutSearch(organism = "Mus musculus", assay_l1 = "Transcriptomic", limit = 10)

# restrict to long-read (PacBio / Oxford Nanopore) studies
SeqoutSearch("liver fibrosis", long_read = TRUE, limit = 10)

# exact case: LINE retrotransposons, not "cell line"
SeqoutSearch("LINE", case_sensitive = TRUE, limit = 10)

## End(Not run)

Description

X becomes counts and obs becomes meta.data. A GSE or GSM accession is resolved through seqout_counts(). Requires SeuratObject.

Usage

seqout_seurat(
  x,
  sample = NULL,
  max_cells = NULL,
  multimodal = TRUE,
  sample_metadata = FALSE,
  ...
)

Seqout2Seurat(
  x,
  sample = NULL,
  max_cells = NULL,
  multimodal = TRUE,
  sample_metadata = FALSE,
  ...
)

Arguments

Value

A Seurat object.

See Also

seqout_matrix() and matrices().

Examples

## Not run: 
obj <- Seqout2Seurat("GSM8994520", min.cells = 3)

obj <- Seqout2Seurat("GSE297547", sample = "GSM8994520")
counts <- SeqoutListCounts("GSE297547")
obj <- Seqout2Seurat(counts, sample = "GSM8994520")

## End(Not run)

Description

Species is inferred from Ensembl-like row identifiers with genevintage, then sex-chromosome markers are obtained from the same annotation registry used by SexSeek. The returned table contains the dual-evidence coordinates (y_percent and x_percent) used by seqout_sex_plot().

Usage

seqout_sex(
  x,
  sample = NULL,
  species = NULL,
  model = "ratio",
  symbol_index = NULL,
  ...
)

SeqoutSex(
  x,
  sample = NULL,
  species = NULL,
  model = "ratio",
  symbol_index = NULL,
  ...
)

Arguments

Value

A data frame of per-unit calls with inferred species and evidence coordinates.

Description

Plot dual sex-chromosome evidence

Usage

seqout_sex_plot(x)

SeqoutSexPlot(x)

Arguments

Value

A ggplot object.

Description

The readable units behind a counts table

Usage

seqout_units(counts, preferred_only = TRUE)

SeqoutUnits(counts, preferred_only = TRUE)

Arguments

Value

A list of units.

Description

data selects on the kind column, which says what a study carries:

Usage

single_cell_studies(
  con = .con(),
  min_evidence = 1,
  data = c("any", "matrix", "fastq", "both"),
  limit = NULL,
  offset = 0
)

SingleCellStudies(
  con = .con(),
  min_evidence = 1,
  data = c("any", "matrix", "fastq", "both"),
  limit = NULL,
  offset = 0
)

Arguments

Details

matrix_and_reads

counts and reads, both confirmed

matrix_reads_unscanned

counts plus linked FASTQ no read-derived call has confirmed yet

matrix_only

counts, no linked reads

reads_only

no parsed matrix

"any" includes reads-only studies and warns when rows lack matrices.

reads_only means Pentimento parsed no matrix. GEO supplementary files may contain a matrix readable by seqout_counts().

Value

A study tibble, one row per study.

Examples

# unrestricted, with matrix warning
sc <- SingleCellStudies(limit = 100)

# parsed matrices
SingleCellStudies(data = "matrix", limit = 100)

# linked reads
SingleCellStudies(data = "fastq", limit = 100)

Description

Study counts per chemistry, organism, tissue, modality, assay category, year, and has_matrix/has_fastq/is_long_read status.

Usage

singlecell_facets(con = .con())

SingleCellFacets(con = .con())

Arguments

Value

A tibble with facet, value and studies columns.

See Also

singlecell_projects().

Examples

f <- SingleCellFacets()
f[f$facet == "chemistry", ]

Description

One row per study with matrix or read-derived single-cell evidence, or a declared single-cell classification: modality, chemistries (read-derived barcode chemistry), cell_or_nucleus (read-derived from intron fraction / mito %, independent of chemistry; NA when never scanned or ambiguous), tissues, organisms, sample count, matrix availability, FASTQ/.sra availability, and whether the study also has long-read sequencing.

Usage

singlecell_projects(
  chemistry = NULL,
  organism = NULL,
  tissue = NULL,
  cell_or_nucleus = NULL,
  perturbation_method = NULL,
  intervention_kind = NULL,
  modality = NULL,
  assay_l1 = NULL,
  year = NULL,
  has_matrix = NULL,
  has_fastq = NULL,
  has_sra = NULL,
  is_long_read = NULL,
  q = NULL,
  sort = "year",
  order = "desc",
  limit = NULL,
  offset = 0,
  con = .con()
)

SingleCellProjects(
  chemistry = NULL,
  organism = NULL,
  tissue = NULL,
  cell_or_nucleus = NULL,
  perturbation_method = NULL,
  intervention_kind = NULL,
  modality = NULL,
  assay_l1 = NULL,
  year = NULL,
  has_matrix = NULL,
  has_fastq = NULL,
  has_sra = NULL,
  is_long_read = NULL,
  q = NULL,
  sort = "year",
  order = "desc",
  limit = NULL,
  offset = 0,
  con = .con()
)

Arguments

Details

perturbation_method is a named method (Perturb-seq, CROP-seq, ECCITE-seq, Mosaic-seq, sci-Plex, CRISPR-detect) mentioned in the study's title or description, not a confirmed design; NA means none was detected. intervention_kind (genetic/chemical) and is_pooled follow from which method matched. Filter values come from singlecell_facets().

Value

A study tibble, with a total attribute for the filtered count before limit cut it.

See Also

singlecell_summary(), singlecell_facets().

Examples

SingleCellProjects(tissue = "liver", has_matrix = TRUE, limit = 100)

Description

Totals for studies with matrix or read-derived single-cell evidence, or a declared single-cell classification; the population singlecell_projects() lists. cells excludes studies whose only matrix is unfiltered (raw 10x barcodes, not real cells).

Usage

singlecell_summary(con = .con())

SingleCellSummary(con = .con())

Arguments

Value

A one-row tibble of totals.

See Also

singlecell_facets() for the filter values, singlecell_projects() for the studies themselves.

Examples

SingleCellSummary()

Description

Study counts per platform, resolution, data availability, organism, tissue, readout assay, cell line, sample type and year. Each facet holds its top 100 values.

Usage

spatial_facets(con = .con())

SpatialFacets(con = .con())

Arguments

Value

A tibble with facet, value and studies columns.

See Also

spatial_projects().

Examples

f <- SpatialFacets()
f[f$facet == "platform", ]

Description

One row per study the modality classifier flags "Spatial Transcriptomics". The platform is named by regex over the study's own text (Visium, Xenium, MERFISH, CosMx, Slide-seq, Slide-seqV2, Curio Seeker, Stereo-seq, GeoMx DSP, DBiT-seq, STARmap, osmFISH, HDST, Tomo-seq, Pixel-seq, Molecular Cartography).

Usage

spatial_projects(
  platform = NULL,
  resolution = NULL,
  technology = NULL,
  data_availability = NULL,
  organism = NULL,
  tissue = NULL,
  readout_assay = NULL,
  cell_line = NULL,
  sample_type = NULL,
  year = NULL,
  has_matrix = NULL,
  has_fastq = NULL,
  is_long_read = NULL,
  q = NULL,
  sort = "year",
  order = "desc",
  limit = NULL,
  offset = 0,
  con = .con()
)

SpatialProjects(
  platform = NULL,
  resolution = NULL,
  technology = NULL,
  data_availability = NULL,
  organism = NULL,
  tissue = NULL,
  readout_assay = NULL,
  cell_line = NULL,
  sample_type = NULL,
  year = NULL,
  has_matrix = NULL,
  has_fastq = NULL,
  is_long_read = NULL,
  q = NULL,
  sort = "year",
  order = "desc",
  limit = NULL,
  offset = 0,
  con = .con()
)

Arguments

Details

platforms is empty when the text never names a platform. A named platform is a text mention: a comparison paper can name one it did not run.

resolution groups platforms into "single-cell" (Xenium, MERFISH, CosMx, seqFISH, STARmap, osmFISH, Molecular Cartography), "spot" (coarser than single-cell: Visium, Slide-seq, Curio Seeker, Stereo-seq, DBiT-seq, HDST, Tomo-seq, Pixel-seq) or "roi" (GeoMx DSP, region-of-interest, not single-cell resolution despite reaching this population). technology is "imaging", "sequencing" or "hybrid" (GeoMx DSP: optical ROI selection, sequencing/nCounter readout); it currently tracks resolution one-to-one but is stored separately. A study naming platforms from two groups (e.g. a GeoMx-vs-Xenium comparison) collapses both columns to whichever group is checked first (single-cell/imaging, then spot/sequencing, then roi/hybrid); platforms still lists every name found. data_availability is "both", "matrix_only", "fastq_only" or "neither", from has_matrix and has_fastq. Filter values come from spatial_facets().

Value

A study tibble, with a total attribute for the filtered count before limit cut it.

See Also

spatial_summary(), spatial_facets(), seqout_counts() to read the matrices of a study with has_matrix.

Examples

# Xenium studies in human
SpatialProjects(platform = "Xenium", organism = "Homo sapiens", limit = 10)

# single-cell-resolution spatial studies with both a matrix and FASTQ
SpatialProjects(
  resolution = "single-cell", data_availability = "both", limit = 10
)

Description

Totals for studies flagged single-cell modality "Spatial Transcriptomics", the same population spatial_projects() lists. studies_named_platform counts studies whose own text names a specific platform (Visium, Xenium, MERFISH, ...); the rest only say "spatial transcriptomics" generically.

Usage

spatial_summary(con = .con())

SpatialSummary(con = .con())

Arguments

Value

A one-row tibble of totals.

See Also

spatial_facets() for the filter values, spatial_projects() for the studies themselves.

Examples

SpatialSummary()

Description

Batches project lookup.

Usage

summaries(accessions, con = .con())

Summaries(accessions, con = .con())

Arguments

Value

A tibble, one row per project.

Examples

Summaries(c("GSE168652", "GSE100379"))

Description

Only GEO, ArrayExpress and DDBJ GEA projects expose this endpoint. Each row's gene_corruption is NULL when the file has not been scanned yet for Excel's gene-symbol-to-date autocorrupt bug (e.g. "SEPT1" -> "1-Sep"). NULL means no data, not a clean result.

Usage

supplementary_files(accession, con = .con())

SupplementaryFiles(accession, con = .con())

Arguments

Value

A tibble with one row per supplementary file: accession, source, url, filename, path, download_command, and gene_corruption (a list-column, each element NULL or a list with n_hits, kinds, examples, checked_at).

Examples

supplementary_files("GSE139368")

Description

registered is FALSE until a remote view is created in DuckDB.

Usage

tables(con = .con())

Tables(con = .con())

Arguments

Value

A tibble with table_name, table_type and registered columns.

Description

Study counts per organism, assay category, archive, journal, submitter country, year, and FASTQ/single-cell/long-read status for studies matching term.