Non-linear Normalization for Non-UMI Single Cell RNA-Seq.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Single-cell RNA sequencing (scRNA-seq) data are susceptible to systematic technical noise.
• This noise arises from variations in mRNA capture, amplification efficiency, and sequencing depth, impacting gene expression measurements.
• Existing normalization methods may not adequately address cell- and gene-specific technical variations.

Purpose of the Study:

• To develop and evaluate a non-linear normalization approach for scRNA-seq data.
• To provide cell- and gene-specific normalization factors to mitigate technical noise.
• To demonstrate the efficacy of the SC2P package in reducing technical variation and bias.

Main Methods:

• Implementation of a non-linear normalization strategy.
• Development of the SC2P software package for scRNA-seq data analysis.
• Comparison of SC2P with existing normalization methods.

Main Results:

• The SC2P normalization method significantly reduces technical variation compared to competing approaches.
• SC2P preserves biological variation while effectively removing technical noise.
• The method corrects biases caused by uneven sequencing depths across cell populations.

Conclusions:

• SC2P offers a robust solution for normalizing scRNA-seq data, particularly for experiments without unique molecular identifiers (UMIs).
• The cell- and gene-specific normalization factors enhance the accuracy of downstream analyses.
• This approach improves the reliability of scRNA-seq data interpretation in the presence of technical noise.