Highly dimensional data generated from bacterial whole-genome sequencing is providing an unprecedented scale of information that requires an appropriate statistical analysis framework to infer biological function from populations of genomes. The application of genome-wide association study (GWAS) methods is an appropriate framework for bacterial population genome analysis that yields a list of candidate genes associated with a phenotype, but it provides an unranked measure of importance. Here, we validated a novel framework to define infection mechanism using the combination of GWAS, machine learning, and bacterial population genomics that ranked allelic variants that accurately identified disease. This approach parsed a dataset of 1.2 million single nucleotide polymorphisms (SNPs) and indels that resulted in an importance ranked list of associated alleles of porA in Campylobacterjejuni using spatiotemporal analysis over 30 years. We validated this approach using previously proven laboratory experimental alleles from an in vivo guinea pig abortion model. This framework, termed µPathML, defined intestinal and extraintestinal groups that have differential allelic porA variants that cause abortion. Divergent variants containing indels that defeated automated annotation were rescued using biological context and knowledge that resulted in defining rare, divergent variants that were maintained in the population over two continents and 30 years. This study defines the capability of machine learning coupled with GWAS and population genomics to simultaneously identify and rank alleles to define their role in infectious disease mechanisms.

The spread of SARS-CoV-2 created a pandemic crisis with > 150,000 cumulative cases in > 65 countries within a few months. The reproductive number (R) is a metric to estimate the transmission of a pathogen during an outbreak. Preliminary published estimates were based on the initial outbreak in China. Whole genome sequences (WGS) analysis found mutational variations in the viral genome; however, previous comparisons failed to show a direct relationship between viral genome diversity, transmission, and the epidemic severity. COVID-19 incidences from different countries were modeled over the epidemic curve. Estimates of the instantaneous R (Wallinga and Teunis method) with a short and standard serial interval were done. WGS were used to determine the populations genomic variation and that underpinned creation of the pathogen genome identity (GENI) score, which was merged with the outbreak curve in four distinct phases. Inference of transmission time was based on a mutation rate of 2 mutations/month. R estimates revealed differences in the transmission and variable infection dynamics between and within outbreak progression for each country examined. Outside China, our R estimates observed propagating dynamics indicating that other countries were poised to move to the takeoff and exponential stages. Population density and local temperatures had no clear relationship to the outbreak progression. Integration of incidence data with the GENI score directly predicted increases in cases as the genome variation increased that led to new variants. Integrating the outbreak curve, dynamic R, and SNP variation found a direct association between increasing cases and transmission genome evolution. By defining the epidemic curve into four stages and integrating the instantaneous country-specific R with the GENI score, we directly connected changes in individual outbreaks based on changes in the virus genome via SNPs. This resulted in the ability to forecast potential increases in cases as well as mutations that may defeat PCR screening and the infection process. By using instantaneous R estimations and WGS, outbreak dynamics were defined to be linked to viral mutations, indicating that WGS, as a surveillance tool, is required to predict shifts in each outbreak that will provide actionable decision making information. Integrating epidemiology with genome sequencing and modeling allows for evidence-based disease outbreak tracking with predictive therapeutically valuable insights in near real time.

Background

Global spread of COVID-19 created an unprecedented infectious disease crisis that progressed to a pandemic with >180,000 cases in >100 countries. Reproductive number (R) is an outbreak metric estimating the transmission of a pathogen. Initial R values were published based on the early outbreak in China with limited number of cases with whole genome sequencing. Initial comparisons failed to show a direct relationship viral genomic diversity and epidemic severity was not established for SARS-Cov-2.

Methods

Each country’s COVID-19 outbreak status was classified according to epicurve stage (index, takeoff, exponential, decline). Instantaneous R estimates (Wallinga and Teunis method) with a short and standard serial interval examined asymptomatic spread. Whole genome sequences were used to quantify the pathogen genome identity score that were used to estimate transmission time and epicurve stage. Transmission time was estimated based on evolutionary rate of 2 mutations/month.

Findings

The country-specific R revealed variable infection dynamics between and within outbreak stages. Outside China, R estimates revealed propagating epidemics poised to move into the takeoff and exponential stages. Population density and local temperatures had variable relationship to the outbreaks. GENI scores differentiated countries in index stage with cryptic transmission. Integration of incidence data with genome variation directly increases in cases with increased genome variation.

Interpretation

R was dynamic for each country and during the outbreak stage. Integrating the outbreak dynamic, dynamic R, and genome variation found a direct association between cases and genome variation. Synergistically, GENI provides an evidence-based transmission metric that can be determined by sequencing the virus from each case. We calculated an instantaneous country-specific R at different stages of outbreaks and formulated a novel metric for infection dynamics using viral genome sequences to capture gaps in untraceable transmission. Integrating epidemiology with genome sequencing allows evidence-based dynamic disease outbreak tracking with predictive evidence.

Funding

Philippine California Advanced Research Institute (Quezon City, Philippines) and the Weimer laboratory.

Research in context

Reproductive number is (R) an epidemiological parameter that defines outbreak transmission dynamics. While early estimates of R exist for COVID-19, the sample size is relatively small (<2000 individuals) taken during the early stages of the disease in China. The outbreak is now a pandemic and a more comprehensive assessment is needed to guide public health efforts in making informed decisions to control regional outbreaks. Commonly, R is computed using a sliding window approach, hence assessment of impact of intervention is more difficult to estimate and often underestimates the dynamic nature of R as the outbreak progresses and expands to different regions of the world. Parallel to epidemiological metrics, pathogen whole genome sequencing is being used to infer transmission dynamics. Viral genome analysis requires expert knowledge in understanding viral genomics that can be integrated with the rapid responses needed for public health to advance outbreak mitigation. This study establishes integrative approaches of genome sequencing with established epidemiological outbreak metrics to provide an easily understandable estimate of transmission dynamics aimed at public health response using evidence-based estimates.

Added value of this study

Estimates of R are dynamic within the progression of the epidemic curve. Using the framework defined in this study with dynamic estimates of R specific to each epicurve stage combined with whole genome sequencing led to creation of a novel metric called GENI (pathogen genome identity) that provides genomic evolution and variation in the context of the outbreak dynamics. The GENI scores were directly linked and proportional to outbreak changes when using disease incidence from epicurve stages (index, takeoff, exponential, and decline). By simulating short and standard (2 day and 7 day, respectively) serial intervals, we calculated instantaneous R followed by a global comparison that was associated with changes in GENI. This approach quantified R values that are impacted by public health intervention to change the outbreak trajectory and were linked to case incidence (i.e. exponential expansion or decelerating) by country. Integrating viral whole genome sequences to estimate GENI we were able to infer circulation time, local transmission, and index case introduction. Systematic integration of viral whole genome sequences with epidemiological parameters resulted in a simplified approach in assessing the status of outbreak that facilitates decisions using evidence from genomics and epidemiology in combination.

Implications of all the available evidence

This study created a framework of evidence-based intervention by integrating whole genome sequencing and epidemiology during the COVID-19 pandemic. Calculating instantaneous R at different stages of the epicurve for different countries provided an evidence-based assessment of control measures as well as the underlying genomic variation globally that changed the outbreak trajectory for all countries examined. Use of the GENI score translates sequencing data into a public health metric that can be directly integrated in epidemiology for outbreak intervention and global preparedness systems.

Hungatella hathewayi has been observed to be a member of the gut microbiome. Unfortunately, little is known about this organism in spite of being associated with human fatalities; it is important to understand virulence mechanisms and epidemiological prospective to cause disease. In this study, a patient with chronic neurologic symptoms presented to the clinic with subsequent isolation of a strain with phenotypic characteristics suggestive of Clostridium difficile. However, whole-genome sequence found the organism to be H. hathewayi. Analysis including publicly available Hungatella genomes found substantial genomic differences as compared to the type strain, indicating this isolate was not C. difficile. We examined the whole-genome of Hungatella species and related genera, using comparative genomics to fully examine species identification and toxin production. Orthogonal phylogenetic using the 16S rRNA gene and entire genome analyses that included genome distance analyses using Genome-to-Genome Distance (GGDC), Average Nucleotide Identity (ANI), and a pan-genome analysis with inclusion of available public genomes determined the speciation to be Hungatella. Two clearly differentiated groups were identified, one including a reference H. hathewayi genome (strain DSM-13,479) and a second group that was determined to be H. effluvii, which included our clinical isolate. Also, some genomes reported as H. hathewayi were found to belong to other genera, including Clostridium and Faecalicatena. We show that the Hungatella species have an open pan-genome reflecting high genomic diversity. This study highlights the importance of correctly assigning taxonomic identification, particularly in disease-associated strains, to better understand virulence and therapeutic options.

Vibrio parahaemolyticus is the most common cause of seafood-borne illness reported in the United States. The draft genomes of 132 North American clinical and oyster V. parahaemolyticus isolates were sequenced to investigate their phylogenetic and biogeographic relationships. The majority of oyster isolate sequence types (STs) were from a single harvest location; however, four were identified from multiple locations. There was population structure along the Gulf and Atlantic Coasts of North America, with what seemed to be a hub of genetic variability along the Gulf Coast, with some of the same STs occurring along the Atlantic Coast and one shared between the coastal waters of the Gulf and those of Washington State. Phylogenetic analyses found nine well-supported clades. Two clades were composed of isolates from both clinical and oyster sources. Four were composed of isolates entirely from clinical sources, and three were entirely from oyster sources. Each single-source clade consisted of one ST. Some human isolates lack tdh, trh, and some type III secretion system (T3SS) genes, which are established virulence genes of V. parahaemolyticus Thus, these genes are not essential for pathogenicity. However, isolates in the monophyletic groups from clinical sources were enriched in several categories of genes compared to those from monophyletic groups of oyster isolates. These functional categories include cell signaling, transport, and metabolism. The identification of genes in these functional categories provides a basis for future in-depth pathogenicity investigations of V. parahaemolyticusIMPORTANCEVibrio parahaemolyticus is the most common cause of seafood-borne illness reported in the United States and is frequently associated with shellfish consumption. This study contributes to our knowledge of the biogeography and functional genomics of this species around North America. STs shared between the Gulf Coast and the Atlantic seaboard as well as Pacific waters suggest possible transport via oceanic currents or large shipping vessels. STs frequently isolated from humans but rarely, if ever, isolated from the environment are likely more competitive in the human gut than other STs. This could be due to additional functional capabilities in areas such as cell signaling, transport, and metabolism, which may give these isolates an advantage in novel nutrient-replete environments such as the human gut.