Classify hyperdiploidy status of multiple myeloma patients using gene expression profiles

Area of Science:

• Oncology
• Genetics
• Bioinformatics

Background:

• Multiple myeloma (MM) is a plasma cell cancer characterized by DNA and chromosome copy number alterations.
• MM is classified into hyperdiploid (HMM) and non-hyperdiploid (NHMM) subtypes, which exhibit distinct prognoses and oncogenic pathways.
• Current subtype classification methods like FISH and microarrays are costly and require significant sample input.

Purpose of the Study:

• To develop and validate a gene expression-based method for classifying HMM and NHMM.
• To investigate the hypothesis that chromosome alterations leave an imprint on gene expression via dosage effects.
• To provide a cost-effective and accessible tool for MM subtype classification.

Main Methods:

• Utilized five MM gene expression datasets with FISH-confirmed HMM status.
• Developed a K-nearest-neighbor (KNN) classification model based on gene expression profiles.
• Validated the KNN model's accuracy on independent test datasets.

Main Results:

• Achieved classification accuracies ranging from 0.83 to 0.88 for HMM and NHMM subtypes.
• Demonstrated that gene expression profiles can effectively distinguish between MM subtypes.
• The study highlights the utility of cancer-specific features and ensemble methods for improved accuracy.

Conclusions:

• A novel, accurate, and cost-effective gene expression-based method for classifying multiple myeloma subtypes (HMM and NHMM) has been developed.
• This classification enables researchers to study MM subtype differences and commonalities using readily available expression data.
• The findings support the use of gene expression profiling for understanding cancer biology and prognosis without additional molecular measurements.