Alevin-fry unlocks rapid, accurate and memory-frugal quantification of single-cell RNA-seq data

Area of Science:

• Genomics
• Computational Biology
• Bioinformatics

Background:

• High-throughput single-cell and single-nucleus RNA sequencing (scRNA-seq and snRNA-seq) technologies generate large, complex datasets.
• Existing computational methods face challenges in accurately and efficiently quantifying this data into usable count matrices.
• These challenges hinder downstream analyses and the extraction of crucial biological insights.

Purpose of the Study:

• To introduce a novel computational framework, alevin-fry, for the accurate and efficient quantification of scRNA-seq and snRNA-seq data.
• To address limitations in memory scalability and false-positive expression rates found in other quantification tools.
• To provide a unified approach for both gene expression quantification and RNA velocity analysis.

Main Methods:

• Development and implementation of the alevin-fry computational framework.
• Benchmarking alevin-fry against existing quantification methods for speed and memory usage.
• Demonstration of alevin-fry's capability to quantify both spliced and unspliced RNA molecules.

Main Results:

• Alevin-fry demonstrates superior speed and reduced memory footprint compared to other accurate quantification tools.
• The framework effectively mitigates memory scalability issues and reduces false-positive expression rates.
• Spliced and unspliced molecule counts for RNA velocity analysis can be seamlessly extracted alongside standard gene expression matrices.

Conclusions:

• Alevin-fry offers a highly efficient and accurate solution for quantifying scRNA-seq and snRNA-seq data.
• The framework overcomes key computational bottlenecks, improving data analysis scalability.
• Alevin-fry provides a versatile tool for both gene expression profiling and advanced RNA velocity studies.