umite: fast quantification of Smart-seq3 libraries with improved UMI retrieval

Area of Science:

• Single-cell RNA sequencing
• Bioinformatics tools
• Computational genomics

Background:

• Commercial single-cell RNA sequencing (scRNA-seq) solutions offer robust Uniquely Identifiable Molecule (UMI) quantification.
• Open-source protocols like Smart-seq3 currently lack comparable UMI quantification support.
• Efficient UMI counting is crucial for accurate gene expression analysis in scRNA-seq.

Purpose of the Study:

• To introduce umite, a novel UMI counting pipeline specifically designed for Smart-seq3.
• To enhance the speed and reduce the memory footprint of UMI quantification for Smart-seq3.
• To improve UMI retrieval rates through mismatch-tolerant detection.

Main Methods:

• Development of the umite pipeline utilizing a Snakemake workflow.
• Implementation of efficient, mismatch-tolerant UMI detection algorithms.
• Benchmarking against existing Smart-seq3 quantification tools using public datasets (GSE207085, GSE270928).

Main Results:

• Umite demonstrates superior performance in runtime, disk usage, and memory footprint compared to current Smart-seq3 quantification tools.
• Mismatch-tolerant UMI detection in umite boosts UMI retrieval by 5-15% in benchmark tests.
• Umite exhibits enhanced scalability for analyzing large single-cell datasets.

Conclusions:

• Umite provides a fast, memory-efficient, and scalable solution for Smart-seq3 UMI quantification.
• The pipeline improves UMI detection accuracy and overall analytical efficiency.
• Umite is readily available as an open-source tool with a Snakemake workflow for seamless integration.