Assigning Transcriptomic Subtypes to Chronic Lymphocytic Leukemia Samples Using Nanopore RNA-Sequencing and...

Area of Science:

• Genomics
• Bioinformatics
• Oncology

Background:

• Massively parallel sequencing advanced chronic lymphocytic leukemia (CLL) diagnostics.
• Illumina platforms are robust but costly; Oxford Nanopore Technologies (ONT) offers a cost-effective alternative, especially for resource-limited settings.
• ONT sequencing requires computational strategies to address lower accuracy and throughput.

Purpose of the Study:

• To characterize the CLL transcriptome landscape using integrated short-read and long-read nanopore sequencing data.
• To identify clinically relevant molecular subtypes of CLL.
• To assign these subtypes to nanopore-sequenced samples using machine learning.

Main Methods:

• Integrated analysis of public Illumina RNA sequencing data (608 CLL samples) and in-house ONT data.
• Transcriptome analysis, gene module identification, and subtype classification using oposSOM and supSOM R packages.
• Machine learning (support vector machine regression) for subtype prediction in nanopore-sequenced samples.

Main Results:

• Identified disruptions in gene modules related to T cell cytotoxicity, immune activation, cell cycle, and splicing in CLL.
• Classified CLL samples into distinct transcriptomic subtypes (e.g., T-cell cytotoxic, immune, proliferative) associated with prognosis.
• Successfully assigned transcriptomic subtypes to nanopore-sequenced patient samples using machine learning.

Conclusions:

• The CLL transcriptome can be parsed into functional modules, revealing molecular subtypes with prognostic and therapeutic implications.
• Integrating ONT sequencing, public data, and machine learning provides a cost-effective method for CLL molecular subtyping and prognostic prediction.
• This approach enhances accessibility to personalized CLL care.