Skip to the content.

Interdisciplinary Machine Learning in Science and Engineering ()

Ulisses Braga-Neto, Department of Electrical & Computer Engineering
Neural Network Prediction of Dengue Fever Severity Based on Genetic Polymorphisms

Abstract

We present a machine learning approach for prediction of dengue fever severity based solely on polymorphisms in host genes implicated in native immunity. Using gene polymorphisms provides a reproducible way to identify subjects that are susceptible to severe dengue fever, even in non-infected individuals. A neural network trained on a cohort of 122 Brazilian dengue patients and controls genotyped for 322 innate immunity single nucleotide polymorphisms (SNPs) produced median values of accuracy greater than 86% in predicting which patients would develop severe dengue fever. At the conclusion of the talk, we will discuss how to apply deep convolutional neural networks to classification of genomic data.