cellatlaser

cellatlaser is an open-source software platform designed to support the construction of cellular atlases from single-cell and spatial transcriptomics data. It provides tools for data integration, cell type annotation, and atlas generation, with an emphasis on reproducibility and interoperability across datasets and analysis pipelines.

Key features of cellatlaser include data import from common single-cell and spatial modalities, preprocessing and normalization, batch-effect correction, dimensionality reduction, and clustering. The platform also includes cell type label transfer using reference atlases, trajectory and lineage inference, and spatial localization of cells within tissue contexts. It supports interactive visualization and scripted workflows, enabling researchers to explore and refine atlas annotations.

Workflow within cellatlaser typically starts with data ingestion and quality control, followed by integration across batches or modalities. Users then perform clustering or annotation against reference labels. The atlas-building module assembles cell types into hierarchical maps, optionally incorporating spatial coordinates, and can export lineage trees and cell-type dictionaries. Reproducible pipelines are supported via configuration files and a workflow language.

Data formats and standards supported by cellatlaser include commonly used formats such as AnnData/h5ad, Loom, and 10x Genomics outputs, as well as spatial data with Visium-like coordinates. It supports standard ontologies and cell-type references, and can export results in standard tabular formats or JSON for integration with downstream tools.

Development and community aspects: cellatlaser is maintained as an open-source project under a permissive license. It hosts documentation, tutorials, and API references, and invites community contributions via version control and issue tracking. The project aims to align with community standards for cell atlas construction and data sharing.