nemoR

nemoR is an R package for discovering, downloading, and loading open-access datasets from the Neuroscience Multi-Omic Archive.

Installation

After CRAN release, install with:

install.packages("nemoR")

For a local development install with browsable vignettes, use:

devtools::install("nemoR", build_vignettes = TRUE)
browseVignettes("nemoR")

The shorter command devtools::install("nemoR") installs the package but does not build vignette HTML by default, so browseVignettes("nemoR") may report no vignettes.

The main fetching workflow is:

library(nemoR)

manifest <- nemo_search_manifest(
  taxon = "house mouse",
  data_type = "counts",
  file_format = "h5ad",
  access = "open",
  max_files = 1
)

nemo_download_plan(manifest)

dry_run <- nemo_fetch(
  destdir = "nemo_downloads",
  taxon = "house mouse",
  data_type = "counts",
  file_format = "h5ad",
  access = "open",
  max_files = 1,
  dry_run = TRUE
)

manifest <- nemo_fetch(
  destdir = "nemo_downloads",
  taxon = "house mouse",
  data_type = "counts",
  file_format = "h5ad",
  access = "open",
  max_files = 1,
  max_size_gb = 2,
  verify_checksum = TRUE
)

nemo_fetch() searches NeMO, builds a manifest, checks file size, downloads to your selected folder, and saves a manifest TSV in that folder. Use dry_run = TRUE to preview the files without downloading them.

The NeMO Assets API is also available for known collection IDs:

collections <- nemo_collections()

files <- nemo_files("nemo:col-rmf5gdy", page_size = 2)
manifest <- nemo_manifest("nemo:col-rmf5gdy", files = files)

Search the NeMO Data Portal by metadata criteria:

results <- nemo_search(
  taxon = "house mouse",
  technique = "10x chromium 3' v2 sequencing",
  file_format = "fastq",
  access = "open",
  target = "files",
  size = 20
)

results
attr(results, "pagination")

List searchable portal fields:

nemo_search_fields()

Discover available facet values before searching or downloading:

nemo_species()
nemo_platforms()
nemo_modalities()
nemo_file_formats()
nemo_brain_regions()
nemo_studies()

Facet helpers accept the same filters as nemo_search():

nemo_platforms(taxon = "house mouse")
nemo_file_formats(taxon = "house mouse", technique = "10x chromium 3' v2 sequencing")
nemo_modalities(file_format = "h5ad", access = "open")

Build a download-ready manifest directly from a search:

manifest <- nemo_search_manifest(
  taxon = "house mouse",
  data_type = "counts",
  file_format = "h5ad",
  access = "open",
  max_files = 1
)

manifest

Inspect the planned download before downloading:

nemo_download_plan(manifest)

The returned manifest records where each file came from and what happened during download. Important columns include:

• download_status: not_downloaded, dry_run, downloaded, skipped_existing, or a failure status.
• checksum_verified: TRUE or FALSE when MD5 verification was requested and the archive provides a checksum; otherwise NA.
• query_parameters: the search settings used to create the manifest.
• manifest_schema_version: the manifest format used by nemoR.

For plain-language walkthroughs, see the package vignettes: vignette("nemoR-quick-start", package = "nemoR") and vignette("nemoR-manifests-and-downloads", package = "nemoR").

For processed public files discovered through the NeMO HTTP browser, create a manifest directly from URLs:

h5ad_url <- paste0(
  "https://data.nemoarchive.org/scorch/grant/U01DA053600_akbarian/",
  "Akbarian/transcriptome/nuclei/10x_v3.1/human/processed/counts/",
  "qc_expr_dat_MHBB_ID__010063.h5ad"
)

manifest <- nemo_manifest_from_urls(
  h5ad_url,
  collection_id = "scorch-akbarian-processed-counts"
)

manifest <- nemo_download(manifest)
sce <- nemo_load(manifest)

Downloads are guarded by max_size_gb = 2 by default to reduce accidental large raw-data downloads. Set verify_checksum = TRUE when you want nemoR to compare downloaded files against archive MD5 checksums where available.

Current modules:

• R/api.R: low-level NeMO API helpers
• R/search.R: dataset and file discovery
• R/manifest.R: reproducible file manifests
• R/download.R: local download and verification helpers
• R/readers.R: file-format readers
• R/objects.R: conversion into R/Bioconductor objects