anndata-data-structure

Annotated data matrices provide the standard data structure for single-cell genomics, efficiently organizing expression data with observation and variable metadata. They facilitate storing cell type information, gene annotations, and dimensionality reductions for large-scale experiments.

You can store single-cell RNA sequencing data by creating an annotated data matrix that pairs expression matrices with observation and variable metadata. This structure natively supports layers, embeddings, and graphs for comprehensive large-scale analysis.

Yes, this annotated data matrix format supports reading and writing both .h5ad and .zarr file formats. These formats are crucial for efficiently handling file I/O for large single-cell genomics experiments.

Yes, this data structure integrates seamlessly with scverse ecosystem tools like scanpy and scvi-tools. It is designed to work within this environment to facilitate subsetting, concatenation, and manipulation of single-cell datasets.

Using an annotated data matrix is the best way to manage large-scale single-cell expression matrices, as it handles data subsetting, concatenation, and manipulation. It securely stores raw counts alongside computed embeddings for later analysis.