Septicemic Melioidosis Detection Using Support Vector Machine with Five Immune Cell Types.

Area of Science:

• Infectious Diseases
• Computational Biology
• Immunology

Background:

• Melioidosis, caused by Burkholderia pseudomallei, is prevalent in tropical regions.
• Septicemic melioidosis is the most lethal form, with a 40% mortality rate.
• Early detection is critical for improving patient outcomes.

Purpose of the Study:

• To develop a novel machine learning-based approach for detecting septicemic melioidosis.
• To utilize immune cell profiling for disease identification.
• To evaluate the performance of support vector machine (SVM) models in diagnosing septicemic melioidosis.

Main Methods:

• Development of multiple SVM models for septicemic melioidosis detection.
• Generation of 19 immune cell-related features using Cell type Identification Estimating Relative Subsets Of RNA Transcripts (CIBERSORT).
• Training and independent evaluation of a binomial SVM model using these features.

Main Results:

• The SVM model demonstrated high performance with sensitivity up to 0.962 and specificity up to 0.979.
• Receiver operating characteristic (ROC) curve analysis showed Area Under Curves (AUCs) ranging from 0.952 to 1.000.
• A concise SVM model using specific immune cell types (CD8+ T cells, resting CD4+ memory T cells, monocytes, M2 macrophages, activated mast cells) achieved sensitivity of 0.976 and specificity of 0.993.

Conclusions:

• The developed SVM model is a robust classification tool for septicemic melioidosis detection.
• This machine learning approach can serve as a complementary diagnostic technique.
• Accurate identification of septicemic melioidosis through immune cell analysis can aid in timely treatment and improve prognosis.