scMultiAnnotator

Description

scMultiAnnotator facilitates single-cell annotation according to its gold standards. This Nextflow-orchestrated pipeline implements three annotation algorithms, each using different methodologies:

1. Manual Annotation: Scanpy plots assist the user in manually annotating cell clusters using custom cell markers.

2. Scoring-Based Method: A Python implementation of ScType, which employs its own database of cell markers.

3. Probabilistic Model: Annotates cell types without the need for reference single-cell data, utilising custom cell markers.

Installation

• Nextflow

  First, install Nextflow in line with its documentation:

  # If `java -version` < 11, reinstall Java
  
  curl -s https://get.sdkman.io | bash
  sdk install java 17.0.10-tem
  java -version
  

  # Install Nextflow
  
  curl -s https://get.nextflow.io | bash
  chmod +x nextflow
  sudo mv nextflow /usr/local/bin
  nextflow info
  

• Docker

  If Docker is not installed yet, follow the steps in this tutorial.

Usage

nextflow run leandrotiburske/scMultiAnnotator \
             --counts ${PWD}/counts \
             --organism mouse \
             --tissue "Immune system" \
             --markers ${PWD}/markers.csv

To see a help message:

nextflow run leandrotiburske/scMultiAnnotator --help

Future improvements

Some new features will be introduced in the future to enhance the functionality of this pipeline, including:

1. Process FASTQ files of multiple samples and perform integration;

2. Include even more annotation methodologies (such as reference-based algorithms).