Investigating the uses of machine learning algorithms to inform risk factor analyses: The example of avian infectious bronchitis virus (IBV) in broiler chickens.

Infectious bronchitis virus (IBV) is a contagious coronavirus causing respiratory and urogenital disease in chickens and is responsible for significant economic losses for both the broiler and table egg layer industries. Despite IBV being regularly monitored using standard epidemiologic surveillance practices, knowledge and evidence of risk factors associated with IBV transmission remain limited. The study objective was to compare risk factor modeling outcomes between a traditional stepwise variable selection approach and a machine learning-based random forest Boruta algorithm using routinely collected IBV antibody titer data from broiler flocks. IBV antibody sampling events (n = 1111) from 166 broiler sites between 2016 and 2021 were accessed. Ninety-two geospatial-related and poultry-density variables were obtained using a geographic information system and data sets from publicly available sources. Seventeen and 27 candidate variables were screened to potentially have an association with elevated IBV antibody titers according to the manual selection and machine learning algorithm, respectively. Selected variables from both methods were further investigated by construction of multivariable generalized mixed logistic regression models. Six variables were shortlisted by both screening methods, which included year, distance to urban areas, main roads, landcover, density of layer sites and year, however, final models for both approaches only shared year as an important predictor. Despite limited significance of clinical outcomes, this work showcases the potential of a novel explorative modeling approach in combination with often unutilized resources such as publicly available geospatial data, surveillance health data and machine learning as potential supplementary tools to investigate risk factors related to infectious diseases.

Space-Time Patterns of Poultry Pathogens in the USA: A Case Study of Ornithobacterium rhinotracheale and Pasteurella multocida in Turkey Populations.

Respiratory infections caused by Ornithobacterium rhinotrachealis (ORT) and Pasteurella multocida (PM) bacteria are significant threats to the poultry industry by causing economic losses and welfare issues. Due to characterization difficulties and underutilization of epidemiological tools, description of the spatio-temporal spread of these diseases in the field is limited. The objectives of this retrospective observational cross-sectional study were to (a) investigate the existence of space-time clusters (hotspots); and (b) investigate the association between genetic similarity and spatial proximity for both pathogens using molecular typing and a recently developed Core-Genome Multilocus Sequencing Typing (cgMLST) scheme. ORT (n = 103) and PM (n = 69) isolates from confirmed disease outbreaks from one commercial company between 2013 and 2021 were obtained from a veterinary diagnostic laboratory, characterized using a cgMLST scheme and visualized using a minimum spanning tree. Spatio-temporal cluster analysis using SaTScanTM and a Spearman's rank correlation were performed to investigate clustering and any association between allelic diversity and geospatial distance. The cgMLST sequencing revealed three allelic clusters for ORT and thirteen clusters for PM. The spatio-temporal analysis revealed two significant clusters for PM, one with a 259.3 km cluster containing six cases between May and July 2018 and a 9 km cluster containing five cases between February 2019 and February 2021. No spatio-temporal clusters were found for ORT. A weak negative correlation between allelic diversity and geospatial distance was observed for ORT (r = -0.04, p < 0.01) and a weak positive correlation was observed for PM (r = 0.11, p < 0.01). This study revealed regional spatio-temporal clusters for PM in commercial turkey sites between 2018 and 2021 and provided additional insight into bacterial strain subgroups and the geographical spread of ORT and PM over time.

The genotyping potential of the Mycoplasma synoviae vlhA gene.

Mycoplasma synoviae (MS) continues to cause significant losses to poultry producers, and studying the epizootiology of infection is an important component of MS control. The partial variable lipoprotein hemagglutinin A (vlhA) gene is the only genomic target identified so far for MS sequence typing. The vlhA gene codes for two variable cell surface proteins, lipoprotein and hemagglutinin, and the proposed mechanism for the variation is gene conversion between a single expressed gene and an array of pseudogenes. The upstream portion of the vlhA gene is present in the genome in a single copy (not present in the pseudogenes), and it is the only part of the gene that can be used for targeted sequence typing. However, the 3' end of the vlhA single copy" as well as this region's discriminatory potential for genotyping purposes has not been established. The purpose of this study was to identify the exact limit and the genotyping potential of the vlhA single copy region. New PCR assays were developed to amplify the entire conserved region and part of the variable region of the vlhA gene. Amplification and sequencing were performed on a variety of MS samples and on in vitro sequential generations of a standard MS strain. Sequence analyses determined the site and composition of the most proximal sequence variation that could be attributed to a gene conversion event, and they predicted the end point of the vlhA single copy region. The results indicated that a currently available "revised Hammond" PCR spans the whole single copy of the vlhA gene and exploits the full genotyping potential of this MS genomic target. In addition, this study allows interesting insight into the gene conversion mechanism of MS and offers the opportunity for further investigation this mechanism in mycoplasmas.