scATAnno: An Automated Cell Type annotation Pipeline for scATAC-seq

scATAnno is a Python-based workflow that automates the annotation of single-cell ATAC-seq (scATAC-seq) data using reference atlases. It directly uses chromatin accessibility peaks as input features. scATAnno utilizes large-scale reference atlases constructed from publicly available datasets and various scATAC-seq technologies to annotate cell types in an unsupervised manner. It incorporates uncertainty score metrics to assess the confidence of cell-type assignments. The workflow has been demonstrated to accurately annotate cell types in different case studies, including PBMC, TNBC, and BCC, and can identify unknown cell populations. Overall, scATAnno expedites our understanding of cellular heterogeneity based on chromatin accessibility profiles in complex biological systems.

• Integration of newly generated scATAC-seq data with reference atlases

• Celltype annotation of single-cell chromatin accessibility profile

• Uncertainty score distribution to assess cell-type assignment

---

If you use this pipeline, please cite the following reference:

Jiang et al., scATAnno: Automated Cell Type Annotation for single-cell ATAC-seq Data

BioRxiv Preprint (https://www.biorxiv.org/content/10.1101/2023.06.01.543296v2)

The documentation is organized into the following sections:

Overview

• About scATAnno
• Installation
  • Install through pip
  • Install through github

Reference Atlases

• Reference Atlas
  • Human scATAC Reference: Zhang, Kai, et al. Cell (2021)
  • PBMC scATAC Reference: Satpathy, Ansuman et al. Nature Biotechnology (2019)
  • BCC Tumor-Infiltrating Lymphocytes (TILs) scATAC Reference:

Workflow Details

• Prepare Reference Atlas
  • Healthy Adult Reference Atlas
  • PBMC Reference Atlas
  • BCC TIL Reference Atlas
• Pipeline Overview
  • Original Input
  • Intermediate Output
  • Final Output

Algorithms

• Dimension reduction
• Cell-type Assignment
  • K-Nearest Neighorbors

Tutorials

• Case Study Datasets
• Tutorials
  • PBMC10K Multiome
  • TNBC sample
  • BCC Multiple Samples