Manages annotated data matrices and metadata for single-cell genomics and high-dimensional biological datasets.
AnnData provides a specialized framework for handling annotated matrices, enabling efficient storage and manipulation of experimental measurements alongside multi-dimensional metadata. It serves as the foundational data structure for the scverse ecosystem, supporting large-scale biological datasets through sparse matrix support and memory-efficient 'backed' modes. This skill guides developers through creating objects, performing high-performance I/O operations with formats like .h5ad and Zarr, and executing complex data manipulations such as concatenation and subsetting for bioinformatics workflows.
主な機能
01Advanced concatenation strategies for combining multiple experimental batches
02Efficient management of X matrices, observation metadata (obs), and variable metadata (var)
03Native support for .h5ad, Zarr, Loom, and 10x Genomics H5 formats
04Seamless integration with Scanpy, scvi-tools, and the wider scverse ecosystem
05Memory-efficient backed mode for processing datasets that exceed RAM capacity
060 GitHub stars
ユースケース
01Preprocessing and quality control of single-cell RNA-seq datasets
02Managing large-scale genomics atlases using sparse matrices and on-disk storage
03Integrating and normalizing multiple biological samples for batch correction
