Explorations in Very Early Prognosis of the Human Immune Response to Influenza

Abstract

We conduct machine learning experiments on time-dependent gene expression measurements associated with the immune response to influenza in humans. We employ three partitions of the two data sets focusing on H1N1 only, H3N2 only and H1N1 and H3N2 combined. From a total set of 1439 known biological pathways, we identify the most discriminatory, potentially capable of providing a very early prognosis of infection, focusing on the time period t ≤ 29 hours post infection. We apply a suite of different machine learning algorithms to these partitions including linear, nonlinear, and sparse support vector machines. In addition, we use artificial neural networks (ANN), k-nearest neighbors and classification on Grassmann manifolds. The cAMP Signaling pathway and the genes PAPSS1 and PAPSS2 appeared to play central role in the very early prognosis problem.

Pathway Expression Analysis

“…In this paper we will run our experiments on a human microarray data set with multiple respiratory viruses and studies, the GSE73072 data set 2 . Previous work on these respiratory virus data used machine learning (ML) models, e.g., neural networks, support vector machines (SVM), centroid encoders (CE), and spectral gene network analysis, to identify discriminatory biomarkers within early shedders challenged with influenza and subsequently classify those subjects [3][4][5][6] .…”