Conference Contributions
Optimizing targeted interventions: Identifying Imported SARS-CoV-2 Lineages and Reconstructing Local Dispersal in Large Phylogenetic Datasets (Poster presented at EEID 2024)
A Comparative Machine Learning Analysis of SARS-CoV-2 Wastewater Surveillance Methods for COVID-19 Case Prediction (Poster presented at ASM Microbe 2024)
This work aimed to explore the predictive capabilities of SARS-CoV-2 wastewater surveillance data produced by varying methodologies to better prioritize wastewater monitoring and predictive efforts. More specifically, both quantifiable (qPCR) and non-quantifiable (presence-absence) SARS-CoV-2 wastewater data were used to train Random Forest models to predict county-level COVID-19 case counts. We also tested the influence of replicate depth (n of extraction replicates) on model performance. Finally, we determined the model with the best accuracy and fit and examined the importance of each feature (wastewater catchment, gene target). These insights could allow for more simplified, rapid approaches for wastewater pathogen detection and case incidence prediction.
Phylodynamics of the H5Nx Clade 2.3.4.4b Avian Influenza Outbreak in Europe and Phylogeographic Diffusion and Host Dynamics of HPAI H5Nx Virus Clade 2.3.4.4b in Europe (Posters presented at EEID & CEIRR 2024)
Integrating Molecular Dynamics and Phylodynamics to Unravel Sialic Acid and HA Interactions Patterns in Avian H5 2.3.4.4b Influenza (Poster presented at CEIRR 2024)
A Preliminary Study on Wastewater-Based Genomic Surveillance for the Early Detection of SARS-CoV-2 and its Distinct Transmission Patterns Across the State of Georgia (GPHL poster 2024)
Harrison Yu, Steven Woods, Carter O’Ryan, Lauren Briggman, Jason Harrison, Amanda Feldpausch, Cristina Meza, Hannah-Leigh Crawford, Tonia Parrott, Arunachalam Ramaiah, Nandhakumar Balakrishnan
Wastewater surveillance offers a non-invasive method to track community-level trends of circulating pathogens, complementing traditional clinical testing. The Georgia Public Health Laboratory, in collaboration with Georgia National Wastewater Surveillance System (NWSS) program, has sequenced SARS-CoV-2 from over 700 wastewater samples collected across 17 treatment facilities in Georgia from January 2023 to July 2024. A Nextflow-based bioinformatics pipeline was developed and used to call variants and obtain lineage abundances. This study identified wastewater-based genomic surveillance of SARS-CoV-2 can provide early detection and reveal transmission patterns, such as the introduction of cases from other states into Georgia.
Validation and Comparison of Bioinformatic Workflows for Routine Genomic Surveillance of Group A Streptococcus in a Public Health Laboratory (GPHL poster 2024)
Tatyana Kiryutina, Arunachalam Ramaiah, Mahalet Bekele, Indira Sawh, Steven Woods, Tonia Parrott, Nandhakumar Balakrishnan
Group A Streptococcal (GAS) infections are caused by Streptococcus pyogenes and are considered a serious public health threat. While Whole-Genome Sequencing (WGS) has been performed in public health laboratories for diagnosis and routine genomic surveillance, there is limited data on bioinformatics pipelines validation for GAS. GPHL aimed to compare three bioinformatics pipelines: FLAQ (Southeast Bioinformatics Regional Resources), PHoeNIx (CDC), and Bactopia (Emory, PGCOE partner), for GAS Multi-locus Sequence Typing (MLST), and genomic characterization to determine their suitability for high-quality results. While a few in-house custom scripts were used to obtain outputs in the desired format, PHoeNIx was facile, automatic, comprehensive, and performed precise quality checks compared to FLAQ and Bactopia. Considering the limitations and advantages of these pipelines, we are evaluating these pipelines on other bacterial species to identify the most suitable bioinformatics pipeline for routine use in public health laboratories for genomic surveillance.
Genomic Analysis of the New Delhi metallo-β-lactamase (blaNDM) Producing Carbapenem-Resistant Enterobacterales Informs the Clonal Transmission of Emerging Lineages in Georgia
Arunachalam Ramaiah, Tatyana Kiryutina, Gebre Tiga, Kaelyn Dugger, Addisalem H Bedada, Steven Woods, Eleen Daley, Tonia Parrott, Nandhakumar Balakrishnan
New Delhi metallo-β-lactamase (blaNDM)-producing carbapenem resistance Enterobacterales (CRE) are more commonly associated with travel, however there is increasing incidence of NDM across healthcare settings in State of Georgia. As part of the Antimicrobial Resistance Laboratory Network (ARLN), antimicrobial surveillance and outbreak investigations are being performed in Georgia to detect the emergence and spread of organisms harboring plasmid and chromosomal genes conferring carbapenem resistance mechanisms. Due to the increase in epidemiologically linked NDM-producing CRE in Georgia, we characterized the genomic diversity of CRE isolates containing the blaNDM gene by whole genome sequencing (WGS). Based on blaNDM genomic characterization, 3 variants were documented including blaNDM-1, blaNDM-5, and blaNDM-7. Most within-species clusters of closely related genomes are predominated by one MLST sequence type. These clusters could be of potential interest for assisting outbreak investigation and understanding clonal dissemination within the facilities.
First report on the emergence of carbapenem resistant Pseudocitrobacter frankii sp. nov., harboring blaNDM-7 in Georgia
Arunachalam Ramaiah, Tatyana Kiryutina, Indira Swah, Gebre Tiga, Addisalem H Bedada, Maranibia Oelemann, Steven Woods, Michael Anderson, JoAnna Wagner, Tonia Parrott, Nandhakumar Balakrishnan
NDM-producing Carbapenem-resistant Enterobacterales (CRE) are more commonly associated with travel, however there is increasing incidence of NDM across healthcare settings in the State of Georgia. In this study, gram-negative, motile, oxidase-negative, catalase-positive, blaNDM-7-positive bacterial strain 2023LY77 was isolated from urine culture of 84-year-old male patient. The strain was characterized by phenotypic, genotypic and WGS methods. MALDI-TOF results included a wide variety of species, within Enterobacterales family. The 16S rRNA and WGS based ANI, in silico-DNA-DNA hybridization (isDDH), phylogenetic and pairwise comparison of this genome with closest type-strain genomes showed that this strain was distinct from other species within the genus Pseudocitrobacter. Based on the distinct differences from their closest species within the genus Pseudocitrobacter, we propose a new name for strain 2023LY77 as “Pseudocitrobacter frankii sp. nov.”
Bayesian Phylodynamics Reveals the Differences in the Transmission and Maintenance of Highly Pathogenic Avian Influenza in Wild Birds and Poultry (Poster presented at EEID 2025)
Xinyi Zhou and Justin Bahl
This study aims to investigate how host ecology and evolutionary pressures shape the transmission dynamics of highly pathogenic avian influenza (HPAI) at the poultry – wild bird interface. We applied a birth-death model to quantify key epidemiological parameters and assessed differences in selection pressure between the two host groups, using outbreaks in Italy as a case study. Future work will extend this framework to a larger study region.
​
Predictive Spatial Modeling of Highly Pathogenic Avian Influenza Transmission Risk in North America (Poster presented at CEIRR 2025)
Mohamed Bakheet, Oluwatosin Babasola, Sachin Subedi, MHM Mubassir, Tanin Rajamand, Christina Naesborg-Nielsen, and Justin Bahl
This study developed a predictive framework for Highly Pathogenic Avian Influenza (HPAI) risk across North America by integrating waterfowl migration data, livestock and human population densities, and environmental covariates into monthly 1 km risk maps. Using ensemble species distribution models (RF, XGBoost, MaxEnt, LightGBM) for 129 waterfowl species, the team derived seasonal richness and a Temporal Co-occurrence Index, then combined these layers with livestock and human data through a calibrated Random Forest model. The maps achieved strong performance (spatial cross-validation AUC = 0.79) and revealed persistent hotspots along the Mississippi and Atlantic flyways. Calibration corrected for overprediction in data-sparse northern regions, yielding well-aligned probabilities with observed outbreaks. Future work will validate against 2025–26 outbreaks, incorporate dynamic predictors such as livestock movements and land-use changes, and link risk surfaces with viral genomic data through an interactive stakeholder dashboard.
Wastewater-Based Surveillance Captures Sars-COV-2 Early Detection, Cryptic Transmission, and Variant Dynamics (Poster presented at EEID 2025)
Gabriella Veytsel, Amanda Howard Sullivan, Leah Lariscy, Megan Lott, Erin Lipp, Travis Glenn, Ludy Carmola, and Justin Bahl
Our study addresses a critical need for evaluation of wastewater genomic surveillance as a population-wide tool to monitor viral evolution and spread. To our knowledge, we present the first study that uses a Bayesian phylodynamic approach to test the utility of sequences obtained from wastewater beyond comparative genomic diversity. Our analyses provide evidence that leveraging the strengths of wastewater surveillance can offer a powerful approach for molecular epidemiology that complements case-based surveillance. We believe our findings will be of broad and immediate interest and may be invaluable for public health decision-making and hospital planning, especially in areas with low clinical testing rates.
Tracing SARS-CoV-2 Clusters Across Local Scales Using Genomic Data (Poster presented at EEID 2025)
Leke Lyu, Mandev Gill, Guppy Stott, Sachin Subedi, Cody Dailey, Gabriella Veytsel, Magdy Alabady, Kayo Fujimoto, Ryker Penn, Pamela Brown, Roger Sealy, Justin Bahl
We present an analytical workflow to trace imported SARS-CoV-2 clusters through communities using large-scale genome datasets. Our approach pinpoints when, where, and how many introductions occurred, while also tracking the circulation of resulting clusters. By incorporating metrics such as the Source Sink Score, Local Import Score, and Persistence Time, our analysis reveals transmission heterogeneity between subregions of the focal area. These insights are essential for monitoring viral introductions, guiding targeted control measures, and enhancing the ability of local responders to address the challenges of current epidemics and future pandemics.
Investigating the Evolutionary Drivers of 2.3.4.4b H5Nx HPAI Spread Across Species and Geographic Regions in Europe (Poster presented at EEID 2025)
Sachin Subedi, MHM Mubassir, Tanin Rajamand, Mohamed Bakheet, Leke Lyu, Oluwatosin Babasola, and Justin Bahl
This study investigated the spread of Highly Pathogenic Avian Influenza (HPAI) H5Nx clade 2.3.4.4b viruses across Europe by analyzing 1,118 genomic sequences collected between 2016 and 2025. Using phylodynamic models, the team identified wetlands as persistent ecological reservoirs and revealed statistically significant viral transmission routes from wetlands to farms, coastal areas and grasslands. Wild birds consistently emerged as the dominant reservoir, driving spillovers into domestic poultry during peak outbreak years. Geographical analyses uncovered temporally shifting dissemination hubs, with GeoClusters Three and Five serving as major regional centers of spread. Live-poultry trade, road-freight volume, vegetation density, rainfall, and proximity to water bodies positively influenced viral transmission, while intensive agricultural land use reduced spread likelihood. These predictors, combined with genomic data, highlighted how ecological habitats, wild bird movements, and human-mediated trade networks jointly shape viral dynamics. The findings provide critical insights for surveillance and control, emphasizing the need for targeted monitoring in wetland-associated bird populations and high-volume trade regions. Future work will extend these analyses to real-time forecasting frameworks linking ecological, trade, and genomic data for proactive outbreak mitigation across Europe.
An Integrative Atlas of Receptor Binding Evolution in U.S. H5N1 2.3.4.4b Viruses (Poster presented at CEIRR 2025)
M H M Mubassir, Sachin Subedi, Tanin Rajamand, Sihua Peng, Pradeep Chopra, Caroline Page, Rajan Kandel, Mohamed Bakheet, Guppy Stott, Cody Dailey, Gabriella Veytsel, Leke Lyu, Oluwatosin Babasola, Christina Næsborg-Nielsen, Ludy Registre Carmola, Gerardus Josephus Boons, Mark Tompkins, Robert J. Woods, and Justin Bahl
This study developed a structure-informed evolutionary framework to track zoonotic risk of H5N1 clade 2.3.4.4b viruses circulating in North America. Drawing on 13,000 DNA sequences and 2,195 non-redundant HA amino acid sequences from U.S. isolates collected between December 2021 and May 2025, the team integrated molecular evolution with structural biology to assess receptor adaptation.Using high-resolution crystal structures of avian and human sialoside complexes (9DIP, 9DIOP), the researchers quantified the energetic impact of all observed amino acid substitutions. These predictions were validated with atomistic molecular dynamics simulations and binding free energy calculations, then projected onto dated phylogenies to visualize evolutionary trajectories in real time.The approach revealed cumulative pathways of receptor adaptation rather than relying on single engineered “switch” mutations, pinpointing viral lineages. By calibrating structural phenotypes with evolutionary dynamics, the framework transforms static surveillance data into dynamic risk maps.
Integrating Cattle Movement Data and Production Networks to Investigate Source–Sink Dynamics of Clade 2.3.4.4b H5N1 in the U.S. (Poster presented at CEIRR 2025 and EEID 2025)
Tanin Rajamand, Sachin Subedi, MHM Mubassir, Michael W. Sanderson, Mohamed Bakheet, Ludy Registre, and Justin Bahl
This study integrates dairy cattle movement networks, land use, poultry and swine production data, and wild bird habitat proximity to assess HPAI risk in cattle and potential cross-species transmission. Phylogeographic analysis of viral sequences tracks the spread of H5N1 among cattle, while comparative modeling identifies the most effective tools for predicting transmission between sectors.Preliminary findings show strong links between cattle movements and viral spread among cattle. Regions where cattle, poultry, and swine production overlap, such as North Carolina, Minnesota, and Texas, face heightened risk. Major cattle importers and exporters, along with wild bird habitats, create conditions favorable to cross-species transmission.By linking animal movement, land use, and genomic data, this research provides key insights for outbreak mitigation and biosecurity. Findings support improved surveillance, targeted interventions, and strengthened policies to reduce zoonotic and pandemic risk.