AtacAnnoR

AtacAnnoR annotates scATAC-seq data by leveraging scRNA-seq reference gene expression profiles and gene activity profiles to assign cell types and interpret chromatin accessibility at single-cell resolution.


Key Features:

  • Two-Round Annotation Method: Employs a two-step annotation process that assigns cell type labels by iteratively comparing gene activity profiles from scATAC-seq with reference scRNA-seq gene expression profiles.
  • Gene Activity Profile Comparison: Uses gene activity profiles derived from scATAC-seq data as the basis for cross-modality comparison to scRNA-seq references.
  • 'Combine and Discard' Strategy: Incorporates a 'Combine and Discard' strategy to integrate annotations from multiple references and manage discrepancies between reference datasets.
  • Unpaired Reference Support: Supports use of unpaired scRNA-seq reference datasets for annotation when paired scRNA-seq and scATAC-seq data are not available.
  • Benchmark Performance: Demonstrated the highest mean accuracy and balanced accuracy in comparative evaluations against six other methods across 11 benchmark datasets.

Scientific Applications:

  • Cell Type Identification: Assigns cell type labels to scATAC-seq profiles to resolve cellular heterogeneity and link chromatin accessibility patterns to cell identity.
  • Comparative Studies with Unpaired Data: Enables comparative analyses across conditions or studies using unpaired scRNA-seq references when direct pairing is not feasible.
  • Interpretation of Chromatin Accessibility: Facilitates interpretation of regulatory landscapes at the single-cell level by providing annotated cell-type contexts for accessibility signals.

Methodology:

Performs a two-round annotation procedure that compares gene activity profiles from scATAC-seq to scRNA-seq reference gene expression profiles and applies a 'Combine and Discard' strategy to integrate multiple references; evaluated against six other methods on 11 benchmark datasets using mean accuracy and balanced accuracy metrics.

Topics

Details

License:
MIT
Cost:
Free of charge
Tool Type:
library
Operating Systems:
Mac, Linux, Windows
Programming Languages:
R
Added:
2/7/2024
Last Updated:
11/24/2024

Operations

Publications

Tian L, Xie Y, Xie Z, Tian J, Tian W. AtacAnnoR: a reference-based annotation tool for single cell ATAC-seq data. Briefings in Bioinformatics. 2023;24(5). doi:10.1093/bib/bbad268. PMID:37497729.

PMID: 37497729
Funding: - National Natural Science Foundation of China: 31871325, 32170667 - National Key Research and Development Program of China: 2021YFC2301500