Added the listPaths() function to list available paths to subdatasets.
This is now referenced by an error message in fetchDataset() when a path= is required but not supplied.
Support complex queries in searchDatasets() via a human-friendly syntax.
Switched to the ArtifactDB representations for the underlying files. This uses language-agnostic formats (e.g., HDF5, JSON) instead of RDS files to store the various parts of each SingleCellExperiment. The user experience should be more or less the same as the datasets are indistinguishable once loaded into memory.
Added the fetchDataset() function, to create SingleCellExperiment objects from the ArtifactDB file representations.
This uses the alabaster.base package to do the loading, with some optional realization of the assays into memory.
Advanced users can achieve faster loading times by keeping the assays as file-backed matrices.
Introduced saveDataset() and related functions to facilitate user uploads of their own datasets.
This is accompanied by some step-by-step instructions in the vignette, plus some maintainer instructions in the README.
Added searchDatasets() to perform text searches on the metadata for each dataset,
using the SQLite database compiled from the gypsum backend where the files are stored.
Updated some datasets to reflect upstream changes (e.g., in ArrayExpress).
Currently, this affects mostly SegerstolpePancreasData(), where ArrayExpress decided to change the names and contents of various column annotations.
Soft-deprecation of some redundant pieces of information in each dataset.
Some examples are the column names of the Zilionis data, which were not unique and had no meaning;
or the symbols of the rowData in the Segerstolpe data, which were redundant with the row names.
These changes will only take effect when fetchDataset() is used directly;
the per-dataset getter functions have appropriate back-compatibility patches to restore this information.
All getters now have a legacy= option to pull RDS files from ExperimentHub instead of the new formats from gypsum.
Fixed phase assignments in the Leng ESC dataset to correctly report G2 phase.
Added the Bacher T cell dataset.
Added the Bhaduri organoid dataset.
Added the Darmanis brain dataset.
Added the Ernst spermatogenesis dataset.
Added the Fletcher olfactory dataset.
Added the Giladi HSC dataset.
Added the He organ atlas dataset.
Added the Jessa brain dataset.
Added the Nowakowski cortex dataset.
Added the Pollen glia dataset.
Added the Zeisel nervous system dataset.
Added the Zhao immune liver dataset.
Added the Zhong prefrontal cortex dataset.
Added the Bunis HSPC dataset (Dan Bunis).
Added the Zilionis lung dataset (Jens Preussner).
Added the Hermann spermatogenesis dataset (Charlotte Soneson).
Added the Mair and Kotliarov PBMC datasets (Stephany Orjuela).
Added the Stoeckius cell hashing dataset.
Added the Wu kidney snRNA-seq dataset.
Added the Hu cortex snRNA-seq dataset.
Added spike-in concentrations to the altExp rowData for various datasets (Alan O'Callaghan).
Removed deprecated inbuilt datasets.
Added the Paul HSC dataset.
Added a dedicated getter for the ERCC concentrations (Alan O'Callaghan).
Added lots of new ExperimentHub datasets, inspired by simpleSingleCell use cases and Martin Hemberg's website.
All outputs are now SingleCellExperiment instances with spike-ins stored as alternative experiments.
Deprecated inbuilt datasets in favor of ExperimentHub equivalents.
New package scRNAseq, providing single-cell RNA sequencing datasets.