AMULET: a novel read count-based method for effective multiplet detection from single nucleus ATAC-seq data.

Area of Science:

• Genomics
• Bioinformatics
• Molecular Biology

Background:

• Single nucleus ATAC-seq (snATAC-seq) is a powerful tool for studying chromatin accessibility at a single-cell resolution.
• Data sparsity and limited dynamic range in snATAC-seq present challenges for accurate multiplet detection.
• Multiplets, or multiple cells captured within a single droplet, can confound downstream analyses.

Purpose of the Study:

• To develop and evaluate a novel computational tool, AMULET (ATAC-seq Multiplet Estimation Tool), for robust multiplet detection in snATAC-seq data.
• To assess the performance of AMULET in terms of precision and recall using both experimental and simulated datasets.
• To determine optimal data quality thresholds for effective multiplet identification using AMULET.

Main Methods:

• AMULET enumerates genomic regions with more than two uniquely aligned reads per nucleus to identify potential multiplets.
• snATAC-seq data were generated from human blood and pancreatic islet samples.
• Multiplet detection performance was evaluated using donor-based multiplexing for precision and simulated multiplets for recall.

Main Results:

• AMULET demonstrates high precision and recall in detecting multiplets in snATAC-seq data.
• The method was validated on human blood and pancreatic islet samples.
• Effective multiplet identification is achieved when a median read depth of 25,000 valid reads per nucleus is reached.

Conclusions:

• AMULET provides an effective computational solution for detecting multiplets in sparse snATAC-seq data.
• The tool offers improved accuracy compared to existing methods.
• Achieving adequate sequencing depth is crucial for maximizing the performance of AMULET in multiplet identification.