Forecasting dominance of SARS-CoV-2 lineages by anomaly detection using deep AutoEncoders (preprint)

The COVID-19 pandemic exemplified the need for a rapid, effective genomic-based surveillance system to predict emerging SARS-CoV-2 variants and lineages. Traditional molecular epidemiology methods, which leverage public health surveillance or integrated sequence data repositories, are able to characterize the evolutionary history of infection waves and genetic evolution but fall short in predicting future outlooks in promptly anticipating viral genetic alterations. To bridge this gap, we introduce a novel Deep learning, autoencoder-based method for anomaly detection in SARS-CoV-2 (DeepAutoCov). Trained and updated on the public global SARS-CoV-2 GISAID database. DeepAutoCov identifies Future Dominant Lineages (FDLs), defined as lineages comprising at least 25% of SARS-CoV-2 genomes added on a given week, on a weekly basis, using the Spike (S) protein. Our algorithm is grounded on anomaly detection via an unsupervised approach, which is necessary given that FDLs can be known only a posteriori (i.e., after they have become dominant). We developed two concurrent approaches (a linear unsupervised and a posteriori supervised) to evaluate DeepAutoCoV performance. DeepAutoCoV identifies FDL, using the spike (S) protein, with a median lead time of 31 weeks on global data and achieves a positive predictive value ~7x better and 23% higher than the other approaches. Furthermore, it predicts vaccine related FDLs up to 17 months in advance. Finally, DeepAutoCoV is not only predictive but also interpretable, since it can pinpoint specific mutations within FDLs, generating hypotheses on the potential increases in virulence or transmissibility of a lineage. By integrating genomic surveillance with artificial intelligence, our work marks a transformative step that may provide valuable insights for the optimization of public health prevention and intervention strategies.

Early emergence phase of SARS-CoV-2 Delta variant in Florida (preprint)

SARS-CoV-2, the causative agent of COVID-19, emerged in late 2020. The highly contagious B.1.617.2 (Delta) Variant of Concern (VOC) was first identified in October 2020 in India and subsequently disseminated worldwide, later becoming the dominant lineage in the U.S. Despite considerable genomic analysis of SARS-CoV-2 in the U.S., several gaps in the understanding of the local dynamics during early introductions remain, which when elucidated could translate the results of viral genomic epidemiology to actionable mitigation efforts. Here, we explore the early emergence of the Delta variant in Florida, U.S. using phylogenetic analysis of representative Florida and globally sampled genomes. We find multiple independent introductions into Florida primarily from North America and Europe, with a minority originating from Asia. These introductions lead to three distinct clades that demonstrated varying relative rates of transmission and possessed five distinct substitutions that were 3-21 times more prevalent in the Florida sample as compared to the global sample. Our results underscore the benefits of routine viral genomic surveillance to monitor epidemic spread and support the need for more comprehensive genomic epidemiology studies of emerging variants. In addition, we provide a model of epidemic spread of newly emerging VOCs that can inform future public health responses.

SARS-CoV-2 Delta vaccine breakthrough transmissibility in Alachua, Florida (preprint)

BackgroundSARS-CoV-2 Delta variant has caused a dramatic resurgence in infections in the United Sates, raising questions regarding potential transmissibility among vaccinated individuals. MethodsBetween October 2020 and July 2021, we sequenced 4,439 SARS-CoV-2 full genomes, 23% of all known infections in Alachua County, Florida, including 109 vaccine breakthrough cases. Univariate and multivariate regression analyses were conducted to evaluate associations between viral load (VL) level and patient characteristics. Contact tracing and phylogenetic analysis were used to investigate direct transmissions involving vaccinated individuals. ResultsThe majority of breakthrough sequences with lineage assignment were classified as Delta variants (74.6%) and occurred, on average, about three months (104 {+/-} 57.5 days) after full vaccination, at the same time (June-July 2021) of Delta variant exponential spread within the county. Six Delta variant transmission pairs between fully vaccinated individuals were identified through contact tracing, three of which were confirmed by phylogenetic analysis. Delta breakthroughs exhibited broad VL values during acute infection (IQR 1.2 - 8.64 Log copies/ml), on average 38% lower than matched unvaccinated patients (3.29 - 10.81 Log copies/ml, p<0.00001). Nevertheless, 49-50% of all breakthroughs, and 56-60% of Delta-infected breakthroughs exhibited VL above the transmissibility threshold (4 Log copies/ml) irrespective of time post vaccination. ConclusionsDelta infection transmissibility and general VL patterns in vaccinated individuals suggest limited levels of sterilizing immunity that need to be considered by public health policies. In particular, ongoing evaluation of vaccine boosters should address whether extra vaccine doses might curb breakthrough contribution to epidemic spread.

SARS-CoV-2 infection of BNT162b2(mRNA)-vaccinated individuals is not restricted to variants of concern or high-risk exposure environments (preprint)

The emergence of SARS-CoV-2 variants of concern (VOC) has raised questions regarding the extent of protection of currently implemented vaccines. Ten "vaccination breakthrough" infections were identified in Alachua County, Florida, among individuals fully vaccinated with the BNT162b2 mRNA vaccine as a result of social or household transmission. Eight individuals presented mild symptoms in the absence of infection with other common respiratory viruses, confirmed using viral genetic sequencing. SARS-CoV-2 genomes were successfully generated for five of the vaccine breakthroughs and 399 individuals in the surrounding area and were included for reference-based phylogenetic investigation. These five individuals were characterized by infection with both VOCs and low-frequency variants present within the surrounding population. Mutations in the Spike protein were consistent with their respective circulating lineages, with the exception of a viable, low-frequency (approximately 1%) B.1.1.7 mutation, which we describe as a mutation of potential concern. The findings indicate that in cases of limited vaccine protection, infection is not restricted to VOCs or high-risk settings, highlighting the critical need for continued testing and monitoring of infection among individuals regardless of vaccination status.

Differing impacts of global and regional responses on SARS-CoV-2 transmission cluster dynamics (preprint)

Although the global response to COVID-19 has not been entirely unified, the opportunity arises to assess the impact of regional public health interventions and to classify strategies according to their outcome. Analysis of genetic sequence data gathered over the course of the pandemic allows us to link the dynamics associated with networks of connected individuals with specific interventions. In this study, clusters of transmission were inferred from a phylogenetic tree representing the relationships of patient sequences sampled from December 30, 2019 to April 17, 2020. Metadata comprising sampling time and location were used to define the global behavior of transmission over this earlier sampling period, but also the involvement of individual regions in transmission cluster dynamics. Results demonstrate a positive impact of international travel restrictions and nationwide lockdowns on global cluster dynamics. However, residual, localized clusters displayed a wide range of estimated initial secondary infection rates, for which uniform public health interventions are unlikely to have sustainable effects. Our findings highlight the presence of so-called "super-spreaders", with the propensity to infect a larger-than-average number of people, in countries, such as the USA, for which additional mitigation efforts targeting events surrounding this type of spread are urgently needed to curb further dissemination of SARS-CoV-2.

Earliest Detection to Date of SARS-CoV-2 in Florida: Identification Together With Influenza Virus on the Main Entry Door of a University Building, February 2020 (preprint)

BackgroundQuestions persist about patterns of initial dissemination of SARS-CoV-2 in the United States in early 2020.MethodsIn February and March, 2020, environmental surface swab samples were collected from the handle of the main entry door of a major university building in Florida, as part of a pilot surveillance project screening for influenza. Samples were taken at the end of regular classroom hours, between the dates of February 1-5 and February 19-March 4, 2020. ResultsInfluenza H1N1pdm09 was isolated from the door handle on four of the 19 days sampled.  Both SARS-CoV-2 and influenza virus were detected in the sample collected on February 21, 2020. Based on sequence analysis, the Florida SARS-CoV-2 strain (designated UF-11) was identical to strains being identified in Washington state during the same time period, while the earliest similar sequences were sampled in China/Hubei between Dec 30th 2019 and Jan 5th 2020. The first human case of COVID-19 was not officially reported in Florida until March 1st. In an analysis of sequences from COVID-19 patients in this region of Florida, there was only limited evidence of subsequent dissemination of the UF-11 strain.  Identical or highly similar strains, possibly related through a common transmission chain, were detected with increasing frequency in Washington state between end of February and beginning of March. ConclusionsOur data provide further documentation of the rapid early spread of SARS-CoV-2, and underscore the likelihood that closely related strains were cryptically circulating in multiple U.S. communities before the first “official” cases were recognized.

A snapshot of SARS-CoV-2 genome availability up to 30th March, 2020 and its implications (preprint)

The SARS-CoV-2 pandemic has been growing exponentially, affecting nearly 900 thousand people and causing enormous distress to economies and societies worldwide. A plethora of analyses based on viral sequences has already been published, in scientific journals as well as through non-peer reviewed channels, to investigate SARS-CoV-2 genetic heterogeneity and spatiotemporal dissemination. We examined full genome sequences currently available to assess the presence of sufficient information for reliable phylogenetic and phylogeographic studies in countries with the highest toll of confirmed cases. Although number of-available full-genomes is growing daily, and the full dataset contains sufficient phylogenetic information that would allow reliable inference of phylogenetic relationships, country-specific SARS-CoV-2 datasets still present severe limitations. Studies assessing within country spread or transmission clusters should be considered preliminary at best, or hypothesis generating. Hence the need for continuing concerted efforts to increase number and quality of the sequences required for robust tracing of the epidemic. Significance StatementAlthough genome sequences of SARS-CoV-2 are growing daily and contain sufficient phylogenetic information, country-specific data still present severe limitations and should be interpreted with caution.

Regaining perspective on SARS-CoV-2 molecular tracing and its implications (preprint)

During the past three months, a new coronavirus (SARS-CoV-2) epidemic has been growing exponentially, affecting over 100 thousand people worldwide, and causing enormous distress to economies and societies of affected countries. A plethora of analyses based on viral sequences has already been published, in scientific journals as well as through non-peer reviewed channels, to investigate SARS-CoV-2 genetic heterogeneity and spatiotemporal dissemination. We examined all full genome sequences currently available to assess the presence of sufficient information for reliable phylogenetic and phylogeographic studies. Our analysis clearly shows severe limitations in the present data, in light of which any finding should be considered, at the very best, preliminary and hypothesis-generating. Hence the need for avoiding stigmatization based on partial information, and for continuing concerted efforts to increase number and quality of the sequences required for robust tracing of the epidemic.