A Novel Method to Detect Bias in Short Read NGS Data

Area of Science:

• Genomics
• Bioinformatics
• Molecular Biology

Background:

• Transcriptomic data analysis requires robust methods to distinguish biological signals from technical biases.
• Sequence-specific biases can confound the interpretation of gene expression levels derived from short-read sequencing.
• Identifying and mitigating these biases is essential for accurate downstream analyses.

Purpose of the Study:

• To develop and present a novel computational method for detecting sequence-specific bias in short-read Next-Generation Sequencing data.
• To validate the method's efficacy using experimental transcriptomic datasets.
• To provide a publicly available software tool for broader application in transcriptomics.

Main Methods:

• The method relies on calculating intra-exon correlations between specific sequence motifs.
• It assumes that short reads originating from the same exon exhibit correlated patterns.
• Implementation was performed using Apache Spark for efficient large-scale data analysis.

Main Results:

• Analysis of Drosophila melanogaster eye-antennal disc datasets revealed significant variations attributable to motif GC content.
• Motif GC content bias was found to be more pronounced than bias related to overall exon GC content.
• The developed software successfully identified sequence-specific biases in the tested datasets.

Conclusions:

• The novel method effectively detects sequence-specific biases in transcriptomic data.
• Motif GC content represents a significant source of bias in short-read sequencing data.
• The software offers a valuable tool for cross-experiment transcriptome data analysis in eukaryotes.