Summary

EcoTyper is a machine learning framework designed for large-scale identification of cell states and cellular ecosystems (ecotypes) from gene expression data. It supports bulk RNA-seq, single-cell RNA-seq (scRNA-seq), and spatial transcriptomics, enabling researchers to discover and characterize complex cellular compositions within tissues. The tool offers both recovery of pre-defined states and de novo discovery, providing a comprehensive approach to cellular ecosystem analysis.

How It Works

EcoTyper employs Nonnegative Matrix Factorization (NMF) as its core algorithm for cell state discovery. For bulk and spatial transcriptomics, it integrates with CIBERSORTx for cell type fraction estimation and expression purification. For scRNA-seq, it adapts NMF to work on correlation matrices derived from cell expression profiles to handle sparsity. The framework also introduces a novel Adaptive False Positive Index (AFI) for quality control of identified cell states.

Quick Start & Requirements

Installation involves cloning the GitHub repository. Key requirements include R (v3.6.0+) and a suite of R packages (ComplexHeatmap, NMF, ggplot2, etc.). Advanced functionalities, particularly for spatial transcriptomics and de novo discovery from bulk data, necessitate Docker or Singularity and CIBERSORTx executables. Users may also require xquartz on macOS. Computationally intensive tasks suggest the use of multi-processor servers or HPC clusters, with recommendations for >32GB RAM for spatial analysis and >50-100GB RAM for scRNA-seq discovery.

Highlighted Details

• Spatial EcoTyper: An extension for spatial transcriptomics data (e.g., Xenium, Atera, MERSCOPE), published in Nature.
• Pre-defined States: Offers pre-identified cell states and ecotypes for carcinomas and diffuse large B cell lymphoma (DLBCL).
• Multi-modal Data Support: Capable of analyzing bulk RNA-seq, scRNA-seq, and spatial transcriptomics data.
• De Novo Discovery: Enables the discovery of novel cell states and ecotypes from user data.
• Quality Control: Implements an Adaptive False Positive Index (AFI) for robust cell state identification.

Maintenance & Community

The project is hosted on GitHub. No specific details regarding active maintenance, community forums (like Discord/Slack), or key contributors were found in the provided README.

Licensing & Compatibility

The license type and any compatibility notes for commercial or closed-source use are not specified in the provided README.

Limitations & Caveats

EcoTyper is computationally demanding, requiring significant RAM and processing power, making it best suited for server or HPC environments. Its setup involves managing multiple R package dependencies and potentially Docker/Singularity containers. The absence of explicit licensing information is a notable adoption blocker.