MACA

MACA: Marker-based automatic cell-type annotation for single cell expression data

1. Installment

MACA works anndata format and is compatible with pipeline analysis through scanpy

 pip install scanpy==1.6.0, anndata==0.7.5, scikit-learn
 
 pip install MACA-Python

2. Tutorial for basic use of MACA

See /Tutorial/Basic use of MACA/

 MACA_tutorial.ipynb

3. Tutorial for integrated annotation

See /Tutorial/Integrated annotation via MACA/

 MACA_integrated_annotation_humanheart.ipynb

 MACA_integrated_annotation_humanpancreas.ipynb

 MACA_integrated_annotation_humanPBMC.ipynb

4. Standardization of cell type annotation across COVID19 datasets via MACA

See /Tutorial/Integrated annotation via MACA/

 MACA_integrated_annotation_COVID19.ipynb

5. Cell-type annotation for 10X Visium data

See /Tutorial/

 MACA_transfer_annotation_spatialbrain10xVisium.ipynb

6. Citation

Xu et al. "MACA: marker-based automatic cell-type annotation for single-cell expression data". Bioinformatics

Update 03/12/2021

MACA was modified for parallel computing. For combined ~647K single nuclei human heart data (Tucker et al, Circulation 2020 and Litviňuková et al, Nature 2020), annotation through MACA takes 24 mins with NMI as 0.739 and ARI as 0.818 against authors' annotations.

Update 11/14/2021

We established a new github repo named MASI (https://github.com/hayatlab/MASI), which combines reference data and MACA for fast label transferring.

Update 03/20/2022

We uploaded ScTypeDB, a combination of PanglaoDB and CellMarker, as cell-type marker database, and tested its performance in cell-type annotation. ScTypeDB is compatible to annotation via MACA.

Ianevski et al. "Fully-automated and ultra-fast cell-type identification using specific marker combinations from single-cell transcriptomic data". Nature Communications

Statement

GPU-supported research has speeded up integrative discoveries across single-cell studies. However, access to a good graphic card for model training is not taken granted, especially in undeveloped and developing countries. Even renting a gpu instance on the cloud is costy for researchers.

We devote to make integrative single-cell analysis accessible for most people, and MACA is a cheap solution to label transferring for large single-cell data. MACA annotates 1 million cells for 40 minutes, on a personal laptop with i7-8550U CPU, 16GB memory, and no GPU support.