Guidelines for poster preparation
Please prepare your poster in A1 portrait format (59cm wide x 84cm long). Do not laminate your poster or use heavy printing material. Further information about poster sizes can be found on the following link:
Posters larger than A1 will only be displayed subject to the availability of space.
Please ensure you have appropriate permissions for the publication of your abstract from the original copyright holders. Should you wish your abstract not to be published, please notify us in writing at the time of abstract submission.
Posters will be displayed for the full duration of the conference.Titles of accepted poster abstracts will be displayed below.
(Presenters in Bold)
If your abstract has been accepted for presentation but it does not appear in the list below, please let us know as soon as possible by email on FluOxford@gmail.com.
Evaluation of follow-up management for H9N2 vaccination in Korea, 2007-2017
Hyun-Kyu Cho, Yong-Myung Kang, Myoung-Heon Lee, Hyun-Mi Kim, Young-Eun Kang and Hyun-Mi Kang
Avian Influenza Research & Diagnostic Division, Animal and Plant Quarantine Agency, 177 Hyeoksin 8-ro, Kimcheon 39660, Republic of Korea
H9N2 low pathogenic avian influenza (LPAI) is a subtype of influenza A virus and found in various species of birds, including poultry and wild birds. The first case of H9N2 LPAI outbreak in Korea was 1996. After that, it has become endemic since 1999 and caused continuous economic losses to poultry industry. The Korean government decided vaccination to minimize economic damage, and the vaccine has been distributed to local farms since the first half of 2007. The present study aimed to evaluate follow-up management of H9N2 vaccination such as sales volume, immunogenicity of commercial vaccines and variation of infection in field. The total sales of H9N2 vaccine was about 670 million doses which drastically increased by 768% in 2017 compared to 2007. Vaccine sales has gradually increased over year but declined at last year, as a result of drastic decrease of H9N2 outbreak by enhanced quarantine measure such as prohibition of livestock transactions in traditional live bird markets. In total sales, combination vaccine was widely used 91%, and 30% of vaccines were distributed free of charge by local governments. In addition, the number of vaccinated farms in each region was proportional to the population of layer and breeding chickens indicating that vaccination mainly focus on long-term breeding poultry. Gyeonggi province has the highest the number of vaccination in layer chicken followed by Gyeongbuk and Chungnam. Chungnam province has the highest in breeding chicken followed by Jeonbuk and Gyeonggi. Meanwhile, the seroprevalence of antibodies for 10% of the whole vaccinated farm was close to the Korean criteria for LPAI vaccine efficacy and there were no cases of infection in field since 2009. Therefore, given these results, H9N2 vaccination has been relatively successful. Furthermore, subsequent H9N2 vaccination need to continuous monitoring and updating the seed strain in consideration of the genetic variation between vaccine and field strain.
Real-time forecasts of seasonal outbreaks for US influenza using Deep learning method
Soo Beom Choi1,2, Insung Ahn1,2
1Dept of Data-centric Problem Solving Research, Korea Inst of Science & Technology Information, Daejeon, Korea
2Center for Convergent Res of Emerging Virus Infection, Korea Res Inst of Chemical Technology, Daejeon, Korea
Previous studies focused on predicting the influenza-like illness (ILI) incidence rate using data from ILI-related queries on Google and other sources. However, studies that predict seasonal influenza epidemics using influenza surveillance data from other countries are rare. This study aimed to investigate the cross-correlation between time series data for ILI, influenza A and B type viruses in US. Along with cross-correlations, we used the Deep learning method to forecast the real-time seasonal influenza outbreaks up to 26 weeks per each week. The FluNet database of the WHO Global Influenza Surveillance Network (URL: http://apps.who.int/flumart/Default?ReportNo=12) contains the following variables reported by 160 countries: number of influenza A and B viruses detected by subtype, number of influenza-positive virus, and so on. We collected surveillance data of 160 countries from year 2010 to 2018. We selected countries with high correlation coefficient and +1 or more week time lag by using the cross-correlation method. Three prediction models for ILI, influenza A and B types were constructed by Deep learning using surveillance data from the selected countries. The hyper-parameters for Deep learning were three of layers, ReLU for activation, and Adam for optimizer. As a result, seasonal patterns for influenza A viruses in Chile were highly correlated with those of the US with 28-week preceding time lag (0.830). The seasonal patterns of influenza B virus in Australia were highly correlated with those of the US with 25-week preceding time lag (0.809). The performances of the Deep learning models to forecast influenza incidences after 26 weeks were acceptable. Our results can help to establish a preventative strategy for seasonal influenza in US by monitoring Chile and Australia. Using our forecasting method, not only the US, but also other countries could make their own forecasting model to make more appropriate and effective interventions in public health department.
Computational Identification of Influenza Neuraminidase Mutations that may improve the fitness of Oseltamivir-resistant A(H1N1)pdm09 viruses
Vithiagaran Gunalan1, Rubaiyea Farrukee2,3, Ding Y Oh2, Ian G Barr2,3, Aeron C Hurt2,3, Sebastian Maurer-Stroh1,4,5
1Bioinformatics Institute, Agency for Science, Technology and Research, Singapore
2WHO Collaborating Centre for Reference and Research on Influenza, Peter Doherty Institute, Victoria, Australia
3University of Melbourne, Department of Microbiology and Immunology, Peter Doherty Institute, Victoria, Australia
4National Public Health Laboratories, Communicable Diseases Division, Ministry of Health, Singapore
5School of Biological Sciences, Nanyang Technological University, Singapore
Neuraminidase Inhibitors (NAIs) specifically inhibit the enzymatic activity of neuraminidase (NA) proteins on influenza A and B viruses, and are currently as the most widely used influenza antivirals. Key resistance mutations in the neuraminidase binding pocket such as H275Y have been characterised, but some of these have been shown to affect NA expression or activity. H275Y-bearing isolates do not seem to be as fit as sensitive ones given the lack of widespread community detection, however secondary mutations such as V241I and N369K have been observed in isolated viruses, suggesting an overall positive effect on the fitness of these isolates, which has been characterised as an effect on NA expression and activity in vitro. In an effort to identify other such permissive mutations, a computational method was used to chart putative pathways based on Gibbs free energy changes by which the accumulation of such mutations could allow for the development of fit resistant strains. Changes in free energy (ΔΔG) of the NA protein were calculated using the FoldX force field, using permutations of mutations representing serial accumulations in NA. From these, 3 candidate permissive mutations were identified for in vitro validation. These were cloned into the NA of A/South Australia/16/2017 with or without H275Y along with background NA mutations observed in sensitive H1N1pdm09 isolates. Mammalian cells transfected with these constructs were assayed for NA expression and activity, and one out of the 3 candidates was seen to exert a positive effect on NA expression, suggesting that it might be able to partially offset the structural instability of NA caused by H275Y. No effect was observed on NA activity, however. This study adds to our understanding of the spectrum of possible mutations that may improve the fitness of H275Y isolates. The effect of this mutation on viral fitness will be evaluated using a ferret model of influenza virus infection and transmission.
Differential interactions between the aconitate decarboxylase (ACOD1) / itaconic acid axis and viral and bacterial pathogens
Dennis Holzwart1, Azeem A. Iqbal1 , Mohamed A. Tantawy1,2, Aaqib Sohail1, Maike Kuhn1, Andrea Kröger4,5, Aurélie Ducroux6 , Christine Goffinet6 , Volkhard Kaever3, Frank Pessler1, 4, 7
1Institute for Experimental Infection Research, TWINCORE, Hannover, Germany
2Hormones Department, Medical Research Division, National Research Center (NRC), Cairo, Egypt
3Research Core Unit Metabolomics, Institute of Pharmacology, Hannover Medical School, Hannover, Germany
4Helmholtz Centre for Infection Research, Braunschweig, Germany
5Institute of Medical Microbiology and Hospital Hygiene, Otto-von-Guericke University, Magdeburg, Germany
6Institute for Experimental Virology, TWINCOR, Hannover, Germany
7Centre for Individualised Infection Medicine, Hannover, Germany
Aconitate decarboxylase 1 (ACOD1) catabolizes the conversion of cis-aconitate to itaconate and exhibits anti-inflammatory and antibacterial effects. It is highly induced during macrophage activation, but little is known about its role in infection of human cells or tissues with different bacterial or viral pathogens. Using the human monocytic leukemia cell line THP-1, which can be differentiated to a macrophage-like phenotype, we have investigated regulation of ACOD1/itaconate by toll-like receptor ligands and differences in the interactions of ACOD1/itaconate with Staphylococcus aureus, Pseudomonas aeruginosa, influenza A virus (IAV), vesicular stomatitis virus, cytomegalovirus, and human immunodeficiency virus type 1. For the TLR ligands the induction of the ACOD1/itaconate axis was highest with lipopolysaccharide, intermediate with Pam3csk and R848, and absent with ODN2216. Accordingly, induction was stronger by P. aeruginosa than by S. aureus, IAV or VSV, while CMV and HIV-1 did not induce ACOD1 expression. Notably, infection with the RNA viruses correlated positively with virulence and interferon induction. In co-infection between bacteria and IAV, the greatest induction of the ACOD1/itaconate response was detected in P. aeruginosa/IAV coinfection. In ACOD1 KO cells (which synthesize only greatly reduced concentrations of itaconate) there was an increased inflammatory response to S. aureus and P. aeruginosa as well as increased intracellular survival of P. aeruginosa, which was reversed by addition of itaconate. Thus, the ACOD1/itaconate response was functionally most significant in Gram-negative infection, particularly during co-infection with IAV. Consequently, infections of these cells with the clinical relevant Gram-negative pathogen Legionella pneumophilia, as well as its co-infection with IAV, are planned to assess the role of the ACOD1/itaconate axis in human host defenses against this pathogen, both in single infection and co-infection with IAV.
Detection of Influenza infection/co-infection through Next Generation Sequence analysis
Bioinformatics Institute, ASTAR, 30 Biopolis Street, #07-01 Matrix, Singapore
Detection and identification of the circulating Influenza strains is an important step in global surveillance to inform the vaccine selection process and for pandemic preparedness. With recent advances made in Next Generation Sequencing (NGS) technology, genome-based identification became much easier. However, it is still hampered by many factors ranging from technical to purely biological. Technical ones worth mentioning include: quality of sample, quality of sequencing, choice of sequencing technology used and last but not least quality and completeness of sequence databases. On the biological side there are three major issues: heavy re-assortment between the genome segments, the possible existence of Influenza as quasispecies and co-infections. To deal with the abovementioned problems, a pipeline was developed with the goals of accurate identification of the major strain/s present in a NGS sample, including quasispecies and detection of possible co-infections.