Machine learning prediction and experimental validation of antigenic drift in H3 influenza A viruses in swine

Abstract

The antigenic diversity of influenza A virus (IAV) circulating in swine challenges the development of effective vaccines, increasing zoonotic threat and pandemic potential. High throughput sequencing technologies are able to quantify IAV genetic diversity, but there are no accurate approaches to adequately describe antigenic phenotypes. This study evaluated an ensemble of non-linear regression models to estimate virus phenotype from genotype. Regression models were trained with a phenotypic dataset of pairwise hemagglutination inhibition (HI) assays, using genetic sequence identity and pairwise amino acid mutations as predictor features. The model identified amino acid identity, ranked the relative importance of mutations in the hemagglutinin (HA) protein, and demonstrated good prediction accuracy. Four previously untested IAV strains were selected to experimentally validate model predictions by HI assays. Error between predicted and measured distances of uncharacterized strains were 0.34, 0.70, 2.19, and 0.17 antigenic units. These empirically trained regression models can be used to estimate antigenic distances between different strains of IAV in swine using sequence data. By ranking the importance of mutations in the HA, we provide criteria for identifying antigenically advanced IAV strains that may not be controlled by existing vaccines and can inform strain updates to vaccines to better control this pathogen.

classLog: Logistic regression for the classification of genetic sequences

“…Logistic regression is a parametric model that performs well on linearly separable classifications. In cases where the data are not linearly separable and that have limited training data, non-parametric models like SVM with an RBF kernel or random forest may perform significantly better, potentially provide easy to understand biological context to feature rankings (Sun et al 2013; Zeller et al 2021), but require more computational time and effort.…”

Evidence of influenza A infection and risk of transmission between pigs and farmworkers

“…Clade classification for each segment was done using the automated classification tool OctoFLU (Chang et al, 2019) (Zeller et al, 2021). The raw assembled sequences were manually sorted into eight FASTA files based on their annotation to the IAV segment.…”

Ecological and evolutionary dynamics of multi-strain RNA viruses

“…Antigenic cartography, a computational technique used for graphical visualization of antigenic distances obtained from inhibition assays 39 , can be used to visualize the genetic and antigenic differences among co-circulating variants and identify clusters of variants with similar immune profiles 40 . Data from panels of cross-reactivity assays can be combined with genetic mapping and epidemiological data and analysed using machine learning and other statistical approaches to identify specific amino acid changes that underlie antigenic phenotypes and potentially result in the emergence of different viral variants [41][42][43] . These tools can be used to refine the relationship between genetic and antigenic variation among co-circulating strains of a virus in a population.…”