zUMIs

zUMIs processes single-cell RNA sequencing (scRNA-seq) data using sample-specific barcodes (BCs) and unique molecular identifiers (UMIs) to produce accurate gene expression counts and mitigate amplification bias.

Key Features:

• Barcode handling: Processes both known and random sample-specific barcodes (BCs).
• UMI collapsing: Collapses UMIs to deduplicate reads, applicable to exon-only or combined exon and intron mapping reads.
• Exon and intron mapping support: Counts reads mapping to exons and optionally to introns to include pre-mRNA signal.
• Intact cell detection: Detects intact cells from the distribution of sequencing reads when BC annotation is missing.
• Adaptive downsampling: Downsamples libraries to manage varying library sizes and to assess whether sequencing has reached saturation.
• Amplification-bias mitigation: Uses UMI-based deduplication to reduce PCR amplification bias in count estimates.

Scientific Applications:

• Single-cell gene expression quantification: Generates accurate per-cell gene expression counts for scRNA-seq analyses.
• Improved gene detection and clustering: Inclusion of intronic reads and UMI handling increases the number of detected genes and can improve cluster resolution.
• Single-nucleus RNA-seq analysis: Applicable to single-nucleus datasets, where over 35% of reads may map to introns, thereby increasing detected genes.
• Library saturation assessment: Uses adaptive downsampling to evaluate sequencing depth and compare libraries of different sizes.

Methodology:

Processes known and random BCs and UMIs, collapses UMIs for exon or exon+intron mapping reads, detects intact cells from read distributions, and performs adaptive downsampling; supports data from scRNA-seq protocols that use BCs and UMIs.