Computational Intelligence applied to Human Genome Data for the Dengue Severity Prognosis

Dengue has become one of the most important worldwide arthropodborne diseases around the world. Here, one hundred and two Brazilian dengue virus (DENV) III patients and controls were genotyped for 322 innate immunity gene loci. All biological data (including age, sex and genome background) were analyzed using Machine Learning techniques to discriminate tendency to severe dengue phenotype development. Our current approach produces median values for accuracy greater than 86%, with sensitivity and specificity over 98% and 51%, respectively. Genome data information from 13 key immune polymorphic SNPs was used under different dominant or recessive models. Our approach is a valuable tool for early diagnosis of the severe form of dengue infection and can be used to identify individuals at high risk of developing this form of the disease even in uninfected individuals. The model also identifies various genes involved dengue severity.