Discovering single nucleotide variants and indels from bulk and single-cell ATAC-seq

Area of Science:

• Genomics
• Epigenetics
• Bioinformatics

Background:

• Whole-genome sequencing (WGS) is the standard for discovering genetic variants in regulatory regions but is costly.
• Most WGS reads originate from non-regulatory genomic areas, limiting efficiency.
• Assay for Transposase-Accessible Chromatin sequencing (ATAC-seq) targets regulatory sequences, suggesting potential for cost-effective variant discovery.

Purpose of the Study:

• To systematically evaluate the utility of ATAC-seq data for discovering regulatory genetic variants.
• To assess the performance of existing variant callers on ATAC-seq data.
• To develop and validate an improved method for variant detection using ATAC-seq.

Main Methods:

• Applied seven established variant callers to bulk and single-cell ATAC-seq datasets.
• Evaluated variant callers' ability to identify single nucleotide variants (SNVs) and insertions/deletions (indels).
• Developed VarCA, an ensemble classifier integrating features from multiple callers to enhance variant prediction.

Main Results:

• The Genome Analysis Toolkit (GATK) showed strong performance among individual callers for SNVs and indels in bulk ATAC-seq data.
• The VarCA ensemble classifier significantly outperformed individual callers on bulk ATAC-seq data for both SNVs and indels.
• VarCA demonstrated high precision and recall for SNVs and indels on single-cell ATAC-seq data, validating its effectiveness across different data types.

Conclusions:

• ATAC-seq data can be effectively utilized for accurate discovery of non-coding regulatory variants without WGS.
• The VarCA ensemble method provides the best overall performance for variant discovery from ATAC-seq data.
• This approach offers a more economical and efficient strategy for identifying regulatory variants.