MetaTiME: Meta-components in Tumor immune MicroEnvironment

MetaTiME learns data-driven, interpretable, and reproducible gene programs by integrating millions of single cells from hundreds of tumor scRNA-seq data. The idea is to learn a map of single-cell space with biologically meaningful directions from large-scale data, which helps understand functional cell states and transfers knowledge to new data analysis. MetaTiME provides pretrained meta-components (MeCs) to automatically annotate fine-grained cell states and plot signature continuum for new single-cells of tumor microenvironment.

Installation

Create a new virtual env and activate (optional)

python -m venv metatime-env; source metatime-env/bin/activate

Use pip to install

pip install metatime

Installation shall be in minutes .

Next we have a tutorial on applying MetaTiME on new TME scRNAseq data to annotate cell states, scoring signature continuum, and test differential signature activity.

Usage

MetaTiME-Annotator

• Use MetaTiME to automatically annotate cell states and map signatures

Interactive tutorial

Use MetaTiME to automatically annotate cell states and map signatures

Method

Reference

Manuscript In Revision. Repo continously being improved! More details will be updated and suggested improvements welcome.

Paper at bioRxiv

Accepted at Nature Communications [Journal Article doi pending]

Training Datasets

Tumor scRNAseq Data for MetaTiME @ Zenodo

• A large collection of uniformly processed tumor single-cell RNA-seq.

• Includes raw data and MetaTiME score for the TME cells.

Dependency

• pandas
• scanpy
• anndata
• matplotlib
• adjustText
• leidenalg
• harmonypy

Dependency version tested:

• pandas==1.1.5
• scanpy==1.8.2
• anndata==0.8.0
• matplotlib==3.5.1
• adjustText==0.7.3
• leidenalg==0.8.3

Contact

Yi Zhang, Ph.D.

yiz [AT] ds.dfci.harvard.edu Twitter | Website Research Fellow Department of Data Science Dana-Farber Cancer Institute Harvard University T.H. Chan School of Public Health