LiBiNorm: an htseq-count analogue with improved normalisation of Smart-seq2 data and library preparation diagnostics

Area of Science:

• Molecular Biology
• Bioinformatics
• Genomics

Background:

• RNA sequencing (RNA-seq) protocols, particularly Smart-seq variations, introduce significant global biases.
• These biases, stemming from enzymatic reactions in cDNA production, cause non-linear, length-dependent RNA over- or under-representation.
• Current RNA-seq software primarily addresses local biases, neglecting global bias correction.

Purpose of the Study:

• To introduce LiBiNorm, a novel command-line program for RNA-seq data analysis.
• To provide diagnostics, quantification, and global bias removal for Smart-seq2 library preparation.
• To enable normalization of gene expression data, correcting for protocol-induced global bias.

Main Methods:

• Development of LiBiNorm, a command-line tool mimicking htseq-count functionality.
• Implementation of global bias removal specific to Smart-seq2 protocols.
• Inclusion of an R script for visualization of normalization results and experimental insights.

Main Results:

• LiBiNorm successfully normalizes gene expression data by correcting for global bias from Smart-seq2.
• The software provides diagnostic data and plots for deeper understanding of library preparation.
• LiBiNorm is identified as the first software to address Smart-seq2 global bias.

Conclusions:

• LiBiNorm offers a crucial solution for accurate gene expression quantification in Smart-seq2 RNA-seq data.
• The tool enhances data reliability by mitigating systematic errors introduced during library preparation.
• LiBiNorm facilitates better biological interpretation of RNA-seq experiments affected by global bias.