Contributions should be made through pull requests. Each PR should consist of a markdown file containing
- a description of the dataset including a link to the appropriate publication or reference.
- direct links to download the count matrix (in the form of an easy-to-load file, like .csv or .tsv files, or
rds
file containing an sparse matrix for R and an AnnDatahdf5
file or amtx
file for python). The most generic file format is .csv/.tsv and the fastest arerds
,npz
orhdf5
. - direct links to download the metadata (in a
csv
with rows indexed by cell names). - bonus: sample loading code for R and python.
An example is datasets/tabula_muris.md
How easy can you make it for someone to get started?