Related Experiment Video
Updated: Aug 14, 2026
14:56
Sample Preparation to Bioinformatics Analysis of DNA Methylation: Association Strategy for Obesity and Related Trait Studies
Published on: May 6, 2022
A systematic benchmarking framework and dual-view optimization strategy for single-cell DNA methylation imputation
Haitian Liang1, Heyang Hua1, Siyu Li1
1School of Mathematical Sciences and LPMC, Nankai University, 94 Weijin Road, Nankai District, Tianjin 300071, China.
Briefings in Bioinformatics
|August 12, 2026
Summary
Accurate single-cell DNA methylation (scDNAm) analysis is challenging due to data sparsity. This study benchmarks imputation methods and introduces BridgeCpG, improving scDNAm data analysis and enabling scalable epigenomic studies.
Area of Science:
- Epigenetics
- Computational Biology
- Genomics
Background:
- Single-cell DNA methylation (scDNAm) profiling offers insights into gene expression regulation.
- Extreme data sparsity in scDNAm data hinders accurate analysis.
- Existing imputation methods lack a comprehensive benchmark for selection.
Purpose of the Study:
- To establish the first systematic benchmarking framework for scDNAm imputation methods.
- To evaluate state-of-the-art imputation methods across diverse datasets and performance dimensions.
- To develop novel strategies to address performance bottlenecks in scDNAm imputation.
Main Methods:
- Developed a systematic benchmarking framework for scDNAm imputation.
- Evaluated five state-of-the-art methods on 13 scDNAm datasets.
- Assessed performance across accuracy, scalability, robustness, and efficiency.
- Proposed a dual-view strategy: BridgeCpG (ensemble model) and adaptive divide-and-conquer (data partitioning).
Main Results:
- Identified critical data attributes and model architectures influencing imputation fidelity.
- Provided quantitative analyses and scenario-aware selection guidelines for imputation methods.
- Demonstrated improved performance through the proposed BridgeCpG and adaptive divide-and-conquer strategies.
- Established a foundation for accurate, high-throughput, and scalable single-cell epigenomic analysis.
Conclusions:
- A rigorous benchmark is crucial for selecting appropriate scDNAm imputation methods.
- The proposed dual-view strategy significantly enhances imputation accuracy and scalability.
- This work provides essential tools and guidelines for advancing single-cell epigenomic research.
