RNAseqCNV: analysis of large-scale copy number variations from RNA-seq data

Area of Science:

• Genomics
• Bioinformatics
• Cancer Research

Background:

• Transcriptome sequencing (RNA-seq) is standard for gene expression and rearrangements in acute lymphoblastic leukemia (ALL).
• The utility of RNA-seq for identifying copy number variations (CNVs) in ALL, a disease with frequent aneuploid subtypes, requires further investigation.
• CNV data is crucial for ALL classification and risk stratification.

Purpose of the Study:

• To introduce RNAseqCNV, a novel method for detecting large-scale CNVs using RNA-seq data.
• To evaluate the accuracy and utility of RNAseqCNV in classifying ALL subtypes.
• To compare RNAseqCNV performance against existing RNA-seq based CNV detection algorithms.

Main Methods:

• Developed RNAseqCNV, a method utilizing normalized gene expression and minor allele frequency models.
• Applied RNAseqCNV to detect large-scale CNVs from RNA-seq data in ALL and acute myeloid leukemia (AML) cohorts.
• Validated CNV calls against DNA-based methods and conventional cytogenetics.

Main Results:

• RNAseqCNV achieved high accuracy in detecting arm-level CNVs in ALL (99.1% overall, 98.3% for non-diploid arms).
• The method demonstrated excellent performance in AML (99.8% overall, 99.4% for non-diploid arms).
• RNAseqCNV outperformed alternative RNA-seq CNV callers, particularly in samples with high CNV burden, and showed high concordance with DNA-based CNV results.

Conclusions:

• RNAseqCNV robustly identifies copy number alterations from RNA-seq data, even without DNA analysis.
• This method significantly enhances the utility of RNA-seq for comprehensive ALL subtyping and risk stratification.
• RNAseqCNV offers a reliable alternative to conventional cytogenetics for CNV detection in leukemia.