The ENCODE Portal is a public repository of experiment records, biosample descriptions, data files, and metadata from the ENCODE Project.
encodeUtils provides an R interface for searching ENCODE metadata, selecting
and verifying files, reading supported local formats, and recording retrieval
provenance. The package is not affiliated with or endorsed by the ENCODE
Project.
See the getting-started vignette for the complete workflow and the function reference for argument and return-value details.
install.packages("BiocManager")
BiocManager::install("encodeUtils")Install the development version with:
install.packages("pak")
pak::pak("ZohebKhan1/encodeUtils")| Function | Purpose |
|---|---|
encode_search() |
Search ENCODE records with exact Portal filters. |
encode_results() |
Extract the main result table. |
encode_list_files() |
List and filter files for experiments. |
encode_file_presets() |
List or inspect file-selection presets. |
encode_select_files() |
Apply file presets and replicate policies. |
encode_download() |
Plan, download, and verify exact files. |
encode_read() |
Read one local file without simplification. |
encode_read_all() |
Read every file in a completed download table. |
encode_prepare_quant() |
Explicitly prepare common quantification tables. |
encode_as_summarized_experiment() |
Combine prepared tables explicitly. |
encode_manifest() |
Record retrieval provenance as JSON. |
ENCODE limits programmatic access to 10 GET requests per second from one user,
group, company, or lab, and warns that abuse can result in an IP block.
encodeUtils uses httr2 to throttle requests to five per second and retry
transient failures.
Downloads are uncapped by default because common FASTQ files exceed hundreds of
megabytes; max_file_size and max_total_size are opt-in guardrails. When no
directory is supplied, files use the package cache from tools::R_user_dir().
An existing destination is reported as "exists" and is not replaced unless
overwrite = TRUE. Transfers use temporary .part files. Size is checked when
reported by ENCODE, and MD5 is checked when ENCODE supplies a checksum.
API responses are not cached. encode_read() limits in-memory reads to 100 MB
by default; use max_size = NULL only when a larger import is intentional.
Large, unsupported, FASTQ, and alignment inputs return an encode_local_file
path object by default so another format-specific tool can process them.
See the ENCODE data-use policy for citation guidance.