Classification of dengue fever patients based on gene expression data using support vector machines

Area of Science:

• Virology
• Immunology
• Bioinformatics

Background:

• Dengue virus (DENV) infection presents as dengue fever (DF) or dengue hemorrhagic fever (DHF).
• The factors determining disease progression remain unclear.
• Host innate immune responses are hypothesized to influence DENV disease outcomes.

Purpose of the Study:

• To explore the role of innate immune gene expression in differentiating DF from DHF.
• To identify potential early prognostic markers for dengue disease severity.
• To apply machine learning for analyzing gene expression patterns in dengue patients.

Main Methods:

• Quantitative real-time PCR (qPCR) was used to measure mRNA expression levels of 12 innate immune response genes.
• Support vector machines (SVM) algorithm was applied to analyze gene expression data from 28 dengue patients (13 DHF, 15 DF).
• Leave-one-out cross-validation was employed to assess model accuracy.

Main Results:

• An SVM model achieved approximately 85% accuracy in classifying DF and DHF patients based on gene expression.
• Seven genes (MYD88, TLR7, TLR3, MDA5, IRF3, IFN-alpha, CLEC5A) were identified as critical for differentiating DF from DHF.
• Incorporating these seven genes increased the SVM model's accuracy to approximately 96%.

Conclusions:

• The study presents a novel application of SVM for classifying dengue disease severity.
• MYD88 and TLR7 were identified as the most significant genes in distinguishing DF from DHF.
• Further research is needed to validate the specific roles of these seven genes in dengue pathogenesis.