MitoSort: Robust Demultiplexing of Pooled Single-cell Genomic Data Using Endogenous Mitochondrial Variants

Area of Science:

• Genomics
• Computational Biology
• Molecular Biology

Background:

• Multiplexing enhances throughput and reduces costs in single-cell genomic studies.
• Current demultiplexing tools rely on nuclear genome variants, limiting their use in methods like scATAC-seq.
• There is a need for robust demultiplexing strategies applicable across diverse single-cell sequencing techniques.

Purpose of the Study:

• To introduce MitoSort, a novel computational tool for sample demultiplexing using mitochondrial germline variants.
• To demonstrate the efficacy of mitochondrial variants as endogenous barcodes for assigning cells to donors.
• To identify cross-genotype doublets in single-cell genomic datasets.

Main Methods:

• Development of MitoSort, a tool leveraging mitochondrial germline variants for demultiplexing.
• Evaluation using in silico pooled mitochondrial scATAC-seq (mtscATAC-seq) libraries.
• Validation with experimentally multiplexed data and a multi-omics sequencing case study.

Main Results:

• MitoSort accurately clusters genotypes and detects doublets in mtscATAC-seq data.
• The tool overcomes limitations of nuclear variant-based methods for scATAC-seq.
• MitoSort shows versatility for various single-cell sequencing approaches with detectable mitochondrial variants.

Conclusions:

• MitoSort provides an accurate and efficient solution for sample demultiplexing in single-cell genomics.
• The tool expands the applicability of multiplexing to sequencing methods with limited nuclear coverage.
• MitoSort facilitates reliable cell origin assignment and doublet detection across diverse single-cell applications.