Datasets and the atlas

U-Chrom keeps no data in its repository. Every dataset it knows reaches you in one call, as a ChromData or as its original files, from three places:

Where

What it is

Browse

Get it

The atlas

public 3-D genome datasets, each one .chromdata.zarr store in object storage (contact maps, RNA, ATAC, images embedded)

the atlas page, the web browser, ds.list_atlas()

ds.load(id) opens it over HTTP; ds.fetch(id) downloads it (.cdz)

Datasets with a loader

published datasets built into a store from their original files the first time you load them

ds.list_datasets() (column load)

ds.load(name)

Original files

the files as their source publishes them (4DN, GEO, Zenodo, GitHub, UCSC): raw formats, reference genomes, annotations

ds.list_datasets()

ds.fetch(name)

 
import uchrom.datasets as ds

ds.list_atlas()                           # the atlas: id, title, study, organism, cells, modalities, sizes
cd = ds.load("stevens2017_mesc")          # an atlas store, opened over HTTP (backed: read in parts)
cd = ds.load("takei2021_mesc")            # a dataset with a loader: built once, then read from disk
path = ds.fetch("takei")                  # an original file: Takei 2021 FOF-CT core table (22 MB, 4DN)

Which one? Analyse a dataset → ds.load (atlas id or loader name: the same call). Work offline or read a whole atlas store many times → ds.fetch(id) once, then ds.load(id) reads the local copy. Read a raw format yourself, or need a genome / annotation file → ds.fetch(name). Your own data → the readers of uchrom.io, saved as a store (data model & storage).

Everything downloaded or built goes to the data directory: $UCHROM_DATA, else ~/.cache/uchrom (ds.data_dir()). Set UCHROM_DATA to keep large data elsewhere or to share a folder of downloads.

Step

Page

API

find a dataset: the atlas page, the web browser, the catalog from Python

Browsing the atlas

ds.list_atlas, chromdata.catalog

a dataset as a ChromData: over HTTP, from a local copy, built by a loader

Loading datasets

ds.load, ds.atlas, ChromData.read(..., backed=True)

files on disk: atlas stores for offline work, original files, reference files

Fetching files

ds.fetch, ds.path, ds.list_datasets, python -m uchrom.datasets

Where every dataset comes from — study, accession, licence, size, how it is fetched or built, who uses it — is listed in data sources.

Tutorial

Guides

API: uchrom.datasets.