Normalizing single-cell RNA sequencing data with internal spike-in-like genes.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Single-cell RNA sequencing (scRNA-seq) requires normalization for accurate analysis.
• Current methods often assume a constant transcriptome, failing to capture drastic cellular changes.
• This limitation impacts the reliability of downstream scRNA-seq analyses.

Purpose of the Study:

• To develop a novel normalization algorithm for scRNA-seq data.
• To address the limitations of whole-transcriptome normalization methods.
• To improve the accuracy and performance of scRNA-seq data analysis.

Main Methods:

• Developed ISnorm (Internal Spike-in-like-genes normalization), an algorithm using constantly expressed genes as internal spike-ins.
• Applied ISnorm to various scRNA-seq datasets exhibiting transcriptome heterogeneity.
• Evaluated the performance of ISnorm in downstream analyses compared to existing methods.

Main Results:

• Demonstrated that single-cell transcriptomes can undergo significant and drastic changes.
• Showcased ISnorm's ability to detect and account for such transcriptome heterogeneity.
• Confirmed that ISnorm normalization improves the performance of downstream scRNA-seq analyses.

Conclusions:

• ISnorm provides a more robust normalization strategy for scRNA-seq data.
• Accounting for transcriptome variability is essential for accurate single-cell analysis.
• The ISnorm algorithm enhances the utility of scRNA-seq for biological discovery.