Understanding early HIV-1 rebound dynamics following antiretroviral therapy interruption: The importance of effector cell expansion.

Most people living with HIV-1 experience rapid viral rebound once antiretroviral therapy is interrupted; however, a small fraction remain in viral remission for an extended duration. Understanding the factors that determine whether viral rebound is likely after treatment interruption can enable the development of optimal treatment regimens and therapeutic interventions to potentially achieve a functional cure for HIV-1. We built upon the theoretical framework proposed by Conway and Perelson to construct dynamic models of virus-immune interactions to study factors that influence viral rebound dynamics. We evaluated these models using viral load data from 24 individuals following antiretroviral therapy interruption. The best-performing model accurately captures the heterogeneity of viral dynamics and highlights the importance of the effector cell expansion rate. Our results show that post-treatment controllers and non-controllers can be distinguished based on the effector cell expansion rate in our models. Furthermore, these results demonstrate the potential of using dynamic models incorporating an effector cell response to understand early viral rebound dynamics post-antiretroviral therapy interruption.

How robust are estimates of key parameters in standard viral dynamic models?

Mathematical models of viral infection have been developed, fitted to data, and provide insight into disease pathogenesis for multiple agents that cause chronic infection, including HIV, hepatitis C, and B virus. However, for agents that cause acute infections or during the acute stage of agents that cause chronic infections, viral load data are often collected after symptoms develop, usually around or after the peak viral load. Consequently, we frequently lack data in the initial phase of viral growth, i.e., when pre-symptomatic transmission events occur. Missing data may make estimating the time of infection, the infectious period, and parameters in viral dynamic models, such as the cell infection rate, difficult. However, having extra information, such as the average time to peak viral load, may improve the robustness of the estimation. Here, we evaluated the robustness of estimates of key model parameters when viral load data prior to the viral load peak is missing, when we know the values of some parameters and/or the time from infection to peak viral load. Although estimates of the time of infection are sensitive to the quality and amount of available data, particularly pre-peak, other parameters important in understanding disease pathogenesis, such as the loss rate of infected cells, are less sensitive. Viral infectivity and the viral production rate are key parameters affecting the robustness of data fits. Fixing their values to literature values can help estimate the remaining model parameters when pre-peak data is missing or limited. We find a lack of data in the pre-peak growth phase underestimates the time to peak viral load by several days, leading to a shorter predicted growth phase. On the other hand, knowing the time of infection (e.g., from epidemiological data) and fixing it results in good estimates of dynamical parameters even in the absence of early data. While we provide ways to approximate model parameters in the absence of early viral load data, our results also suggest that these data, when available, are needed to estimate model parameters more precisely.

Modeling the emergence of viral resistance for SARS-CoV-2 during treatment with an anti-spike monoclonal antibody.

To mitigate the loss of lives during the COVID-19 pandemic, emergency use authorization was given to several anti-SARS-CoV-2 monoclonal antibody (mAb) therapies for the treatment of mild-to-moderate COVID-19 in patients with a high risk of progressing to severe disease. Monoclonal antibodies used to treat SARS-CoV-2 target the spike protein of the virus and block its ability to enter and infect target cells. Monoclonal antibody therapy can thus accelerate the decline in viral load and lower hospitalization rates among high-risk patients with variants susceptible to mAb therapy. However, viral resistance has been observed, in some cases leading to a transient viral rebound that can be as large as 3-4 orders of magnitude. As mAbs represent a proven treatment choice for SARS-CoV-2 and other viral infections, evaluation of treatment-emergent mAb resistance can help uncover underlying pathobiology of SARS-CoV-2 infection and may also help in the development of the next generation of mAb therapies. Although resistance can be expected, the large rebounds observed are much more difficult to explain. We hypothesize replenishment of target cells is necessary to generate the high transient viral rebound. Thus, we formulated two models with different mechanisms for target cell replenishment (homeostatic proliferation and return from an innate immune response antiviral state) and fit them to data from persons with SARS-CoV-2 treated with a mAb. We showed that both models can explain the emergence of resistant virus associated with high transient viral rebounds. We found that variations in the target cell supply rate and adaptive immunity parameters have a strong impact on the magnitude or observability of the viral rebound associated with the emergence of resistant virus. Both variations in target cell supply rate and adaptive immunity parameters may explain why only some individuals develop observable transient resistant viral rebound. Our study highlights the conditions that can lead to resistance and subsequent viral rebound in mAb treatments during acute infection.

Stoichiometry for entry and binding properties of the Env protein of R5 T cell-tropic HIV-1 and its evolutionary variant of macrophage-tropic HIV-1.

Human immunodeficiency virus type 1 typically requires a high density of CD4 for efficient entry as a mechanism to target CD4+ T cells (T-tropic), with CCR5 being used most often as the coreceptor. When target T cells are limiting, the virus can evolve to infect cells with a low density of CD4 such as macrophages (M-tropic). The entry phenotype is known to be encoded in the viral Env protein on the surface of the virus particle. Using data showing a dose response for infectivity based on CD4 surface density, we built a model consistent with T-tropic viruses requiring multiple CD4 molecules to mediate infection, whereas M-tropic viruses can infect cells using a single CD4 receptor molecule interaction. We also found that T-tropic viruses bound to the surface of cells with a low density of CD4 are released more slowly than M-tropic viruses which we modeled to be due to multiple interactions of the T-tropic virus with multiple CD4 molecules to allow the initial stable binding. Finally, we found that some M-tropic Env proteins, as the gp120 subunit, possess an enhanced affinity for CD4 compared with their T-tropic pair, indicating that the evolution of macrophage tropism can be reflected both in the closed Env trimer conformation on the virion surface and, in some cases, also in the open confirmation of gp120 Env. Collectively, these studies reveal differences in the stoichiometry of interaction of T-tropic and M-tropic viruses with CD4 and start to identify the basis of binding differences at the biochemical level. IMPORTANCE: Human immunodeficiency virus type 1 normally targets CD4+ T cells for viral replication. When T cells are limiting, the virus can evolve to infect myeloid cells. The evolutionary step involves a change from requiring a high surface density of CD4 for entry to being able to infect cells with a low density of CD4, as is found on myeloid lineage cells such as macrophage and microglia. Viruses able to infect macrophages efficiently are most often found in the CNS late in the disease course, and such viruses may contribute to neurocognitive impairment. Here, we examine the CD4 binding properties of the viral Env protein to explore these two different entry phenotypes.

The Effects of Human Immunodeficiency Virus Type 1 (HIV-1) Antigen-Expanded Specific T-Cell Therapy and Vorinostat on Persistent HIV-1 Infection in People With HIV on Antiretroviral Therapy.

BACKGROUND: The histone deacetylase inhibitor vorinostat (VOR) can reverse human immunodeficiency virus type 1 (HIV-1) latency in vivo and allow T cells to clear infected cells in vitro. HIV-specific T cells (HXTCs) can be expanded ex vivo and have been safely administered to people with HIV (PWH) on antiretroviral therapy. METHODS: Six PWH received infusions of 2 × 107 HXTCs/m² with VOR 400 mg, and 3 PWH received infusions of 10 × 107 HXTCs/m² with VOR. The frequency of persistent HIV by multiple assays including quantitative viral outgrowth assay (QVOA) of resting CD4+ T cells was measured before and after study therapy. RESULTS: VOR and HXTCs were safe, and biomarkers of serial VOR effect were detected, but enhanced antiviral activity in circulating cells was not evident. After 2 × 107 HXTCs/m² with VOR, 1 of 6 PWH exhibited a decrease in QVOA, and all 3 PWH exhibited such declines after 10 × 107 HXTCs/m² and VOR. However, most declines did not exceed the 6-fold threshold needed to definitively attribute decline to the study intervention. CONCLUSIONS: These modest effects provide support for the strategy of HIV latency reversal and reservoir clearance, but more effective interventions are needed to yield the profound depletion of persistent HIV likely to yield clinical benefit. Clinical Trials Registration. NCT03212989.

Predicting Impacts of Contact Tracing on Epidemiological Inference from Phylogenetic Data.

Robust sampling methods are foundational to many inference problems in the phylodynamic field, yet the impact of using contact tracing, a type of non-uniform sampling used in public health applications, is not well understood. To investigate and quantify how this non-uniform sampling method influences recovered phylogenetic tree structure, we developed a new simulation tool called SEEPS (Sequence Evolution and Epidemiological Process Simulator) that allows for the simulation of contact tracing and the resulting transmission tree, pathogen phylogeny, and corresponding virus genetic sequences. Importantly, SEEPS takes within-host evolution into account when generating pathogen phylogenies and sequences from transmission histories. Using SEEPS, we demonstrate that contact tracing can significantly impact the structure of the resulting tree as described by popular tree statistics. Contact tracing generates phylogenies that are less balanced than the underlying transmission process, less representative of the larger epidemiological process, and affects the internal/external branch length ratios that characterize specific epidemiological scenarios. We also examine a 2007-2008 Swedish HIV-1 outbreak and the broader 1998-2010 European HIV-1 epidemic to highlight the differences in contact tracing and expected phylogenies. Aided by SEEPS, we show that the Swedish outbreak was strongly influenced by contact tracing even after downsampling, while the broader European Union epidemic showed little evidence of universal contact tracing, agreeing with the known epidemiological information about sampling and spread. SEEPS is available at github.com/MolEvolEpid/SEEPS.

AZD5582 plus SIV-specific antibodies reduce lymph node viral reservoirs in antiretroviral therapy-suppressed macaques.

The main barrier to HIV cure is a persistent reservoir of latently infected CD4+ T cells harboring replication-competent provirus that fuels rebound viremia upon antiretroviral therapy (ART) interruption. A leading approach to target this reservoir involves agents that reactivate latent HIV proviruses followed by direct clearance of cells expressing induced viral antigens by immune effector cells and immunotherapeutics. We previously showed that AZD5582, an antagonist of inhibitor of apoptosis proteins and mimetic of the second mitochondrial-derived activator of caspases (IAPi/SMACm), induces systemic reversal of HIV/SIV latency but with no reduction in size of the viral reservoir. In this study, we investigated the effects of AZD5582 in combination with four SIV Env-specific Rhesus monoclonal antibodies (RhmAbs) ± N-803 (an IL-15 superagonist) in SIV-infected, ART-suppressed rhesus macaques. Here we confirm the efficacy of AZD5582 in inducing SIV reactivation, demonstrate enhancement of latency reversal when AZD5582 is used in combination with N-803 and show a reduction in total and replication-competent SIV-DNA in lymph-node-derived CD4+ T cells in macaques treated with AZD5582 + RhmAbs. Further exploration of this therapeutic approach may contribute to the goal of achieving an HIV cure.

Modeling the emergence of viral resistance for SARS-CoV-2 during treatment with an anti-spike monoclonal antibody.

The COVID-19 pandemic has led to over 760 million cases and 6.9 million deaths worldwide. To mitigate the loss of lives, emergency use authorization was given to several anti-SARS-CoV-2 monoclonal antibody (mAb) therapies for the treatment of mild-to-moderate COVID-19 in patients with a high risk of progressing to severe disease. Monoclonal antibodies used to treat SARS-CoV-2 target the spike protein of the virus and block its ability to enter and infect target cells. Monoclonal antibody therapy can thus accelerate the decline in viral load and lower hospitalization rates among high-risk patients with susceptible variants. However, viral resistance has been observed, in some cases leading to a transient viral rebound that can be as large as 3-4 orders of magnitude. As mAbs represent a proven treatment choice for SARS-CoV-2 and other viral infections, evaluation of treatment-emergent mAb resistance can help uncover underlying pathobiology of SARS-CoV-2 infection and may also help in the development of the next generation of mAb therapies. Although resistance can be expected, the large rebounds observed are much more difficult to explain. We hypothesize replenishment of target cells is necessary to generate the high transient viral rebound. Thus, we formulated two models with different mechanisms for target cell replenishment (homeostatic proliferation and return from an innate immune response anti-viral state) and fit them to data from persons with SARS-CoV-2 treated with a mAb. We showed that both models can explain the emergence of resistant virus associated with high transient viral rebounds. We found that variations in the target cell supply rate and adaptive immunity parameters have a strong impact on the magnitude or observability of the viral rebound associated with the emergence of resistant virus. Both variations in target cell supply rate and adaptive immunity parameters may explain why only some individuals develop observable transient resistant viral rebound. Our study highlights the conditions that can lead to resistance and subsequent viral rebound in mAb treatments during acute infection.

Semi-infectious particles contribute substantially to influenza virus within-host dynamics when infection is dominated by spatial structure.

Influenza is an ribonucleic acid virus with a genome that comprises eight segments. Experiments show that the vast majority of virions fail to express one or more gene segments and thus cannot cause a productive infection on their own. These particles, called semi-infectious particles (SIPs), can induce virion production through complementation when multiple SIPs are present in an infected cell. Previous within-host influenza models did not explicitly consider SIPs and largely ignore the potential effects of coinfection during virus infection. Here, we constructed and analyzed two distinct models explicitly keeping track of SIPs and coinfection: one without spatial structure and the other implicitly considering spatial structure. While the model without spatial structure fails to reproduce key aspects of within-host influenza virus dynamics, we found that the model implicitly considering the spatial structure of the infection process makes predictions that are consistent with biological observations, highlighting the crucial role that spatial structure plays during an influenza infection. This model predicts two phases of viral growth prior to the viral peak: a first phase driven by fully infectious particles at the initiation of infection followed by a second phase largely driven by coinfections of fully infectious particles and SIPs. Fitting this model to two sets of data, we show that SIPs can contribute substantially to viral load during infection. Overall, the model provides a new interpretation of the in vivo exponential viral growth observed in experiments and a mechanistic explanation for why the production of large numbers of SIPs does not strongly impede viral growth. Being simple and predictive, our model framework serves as a useful tool to understand coinfection dynamics in spatially structured acute viral infections.

Making waves: Integrating wastewater surveillance with dynamic modeling to track and predict viral outbreaks.

Wastewater surveillance has proved to be a valuable tool to track the COVID-19 pandemic. However, most studies using wastewater surveillance data revolve around establishing correlations and lead time relative to reported case data. In this perspective, we advocate for the integration of wastewater surveillance data with dynamic within-host and between-host models to better understand, monitor, and predict viral disease outbreaks. Dynamic models overcome emblematic difficulties of using wastewater surveillance data such as establishing the temporal viral shedding profile. Complementarily, wastewater surveillance data bypasses the issues of time lag and underreporting in clinical case report data, thus enhancing the utility and applicability of dynamic models. The integration of wastewater surveillance data with dynamic models can enhance real-time tracking and prevalence estimation, forecast viral transmission and intervention effectiveness, and most importantly, provide a mechanistic understanding of infectious disease dynamics and the driving factors. Dynamic modeling of wastewater surveillance data will advance the development of a predictive and responsive monitoring system to improve pandemic preparedness and population health.

Swift and extensive Omicron outbreak in China after sudden exit from 'zero-COVID' policy.

In late 2022, China transitioned from a strict 'zero-COVID' policy to rapidly abandoning nearly all interventions and data reporting. This raised great concern about the presumably-rapid but unreported spread of the SARS-CoV-2 Omicron variant in a very large population of very low pre-existing immunity. By modeling a combination of case count and survey data, we show that Omicron spread extremely rapidly, at a rate of 0.42/day (95% credibility interval: [0.35, 0.51]/day), translating to an epidemic doubling time of 1.6 days ([1.6, 2.0] days) after the full exit from zero-COVID on Dec. 7, 2022. Consequently, we estimate that the vast majority of the population (97% [95%, 99%], sensitivity analysis lower limit of 90%) was infected during December, with the nation-wide epidemic peaking on Dec. 23. Overall, our results highlight the extremely high transmissibility of the variant and the importance of proper design of intervention exit strategies to avoid large infection waves.

Prolonged viral shedding from noninfectious individuals confounds wastewater-based epidemiology.

Wastewater surveillance has been widely used to track and estimate SARS-CoV-2 incidence. While both infectious and recovered individuals shed virus into wastewater, epidemiological inferences using wastewater often only consider the viral contribution from the former group. Yet, the persistent shedding in the latter group could confound wastewater-based epidemiological inference, especially during the late stage of an outbreak when the recovered population outnumbers the infectious population. To determine the impact of recovered individuals' viral shedding on the utility of wastewater surveillance, we develop a quantitative framework that incorporates population-level viral shedding dynamics, measured viral RNA in wastewater, and an epidemic dynamic model. We find that the viral shedding from the recovered population can become higher than the infectious population after the transmission peak, which leads to a decrease in the correlation between wastewater viral RNA and case report data. Furthermore, the inclusion of recovered individuals' viral shedding into the model predicts earlier transmission dynamics and slower decreasing trends in wastewater viral RNA. The prolonged viral shedding also induces a potential delay in the detection of new variants due to the time needed to generate enough new cases for a significant viral signal in an environment dominated by virus shed by the recovered population. This effect is most prominent toward the end of an outbreak and is greatly affected by both the recovered individuals' shedding rate and shedding duration. Our results suggest that the inclusion of viral shedding from non-infectious recovered individuals into wastewater surveillance research is important for precision epidemiology.

Global estimates of the fitness advantage of SARS-CoV-2 variant Omicron.

New variants of SARS-CoV-2 show remarkable heterogeneity in their relative fitness over both time and space. In this paper we extend the tools available for estimating the selection strength for new SARS-CoV-2 variants to a hierarchical, mixed-effects, renewal equation model. This formulation allows us to estimate selection effects at the global level while incorporating both measured and unmeasured heterogeneity among countries. Applying this model to the spread of Omicron in forty countries, we find evidence for very strong but very heterogeneous selection effects. To test whether this heterogeneity is explained by differences in the immune landscape, we considered several measures of vaccination rates and recent population-level infection as covariates, finding moderately strong, statistically significant effects. We also found a significant positive correlation between the selection advantage of Delta and Omicron at the country level, suggesting that other region-specific explanatory variables of fitness differences do exist. Our method is implemented in the Stan programming language, can be run on standard consumer-grade computing resources, and will be straightforward to apply to future variants.

A deep learning approach to real-time HIV outbreak detection using genetic data.

Pathogen genomic sequence data are increasingly made available for epidemiological monitoring. A main interest is to identify and assess the potential of infectious disease outbreaks. While popular methods to analyze sequence data often involve phylogenetic tree inference, they are vulnerable to errors from recombination and impose a high computational cost, making it difficult to obtain real-time results when the number of sequences is in or above the thousands. Here, we propose an alternative strategy to outbreak detection using genomic data based on deep learning methods developed for image classification. The key idea is to use a pairwise genetic distance matrix calculated from viral sequences as an image, and develop convolutional neutral network (CNN) models to classify areas of the images that show signatures of active outbreak, leading to identification of subsets of sequences taken from an active outbreak. We showed that our method is efficient in finding HIV-1 outbreaks with R0 ≥ 2.5, and overall a specificity exceeding 98% and sensitivity better than 92%. We validated our approach using data from HIV-1 CRF01 in Europe, containing both endemic sequences and a well-known dual outbreak in intravenous drug users. Our model accurately identified known outbreak sequences in the background of slower spreading HIV. Importantly, we detected both outbreaks early on, before they were over, implying that had this method been applied in real-time as data became available, one would have been able to intervene and possibly prevent the extent of these outbreaks. This approach is scalable to processing hundreds of thousands of sequences, making it useful for current and future real-time epidemiological investigations, including public health monitoring using large databases and especially for rapid outbreak identification.

Emergence of SARS-CoV-2 escape mutations during Bamlanivimab therapy in a phase II randomized clinical trial.

SARS-CoV-2 mutations that cause resistance to monoclonal antibody (mAb) therapy have been reported. However, it remains unclear whether in vivo emergence of SARS-CoV-2 resistance mutations alters viral replication dynamics or therapeutic efficacy in the immune-competent population. As part of the ACTIV-2/A5401 randomized clinical trial (NCT04518410), non-hospitalized participants with symptomatic SARS-CoV-2 infection were given bamlanivimab (700 mg or 7,000 mg) or placebo treatment. Here¸ we report that treatment-emergent resistance mutations [detected through targeted Spike (S) gene next-generation sequencing] were significantly more likely to be detected after bamlanivimab 700 mg treatment compared with the placebo group (7% of 111 vs 0% of 112 participants, P = 0.003). No treatment-emergent resistance mutations among the 48 participants who received 7,000 mg bamlanivimab were recorded. Participants in which emerging mAb resistant virus mutations were identified showed significantly higher pretreatment nasopharyngeal and anterior nasal viral loads. Daily respiratory tract viral sampling through study day 14 showed the dynamic nature of in vivo SARS-CoV-2 infection and indicated a rapid and sustained viral rebound after the emergence of resistance mutations. Participants with emerging bamlanivimab resistance often accumulated additional polymorphisms found in current variants of concern/interest that are associated with immune escape. These results highlight the potential for rapid emergence of resistance during mAb monotherapy treatment that results in prolonged high-level respiratory tract viral loads. Assessment of viral resistance should be prioritized during the development and clinical implementation of antiviral treatments for COVID-19.

A simple model of COVID-19 explains disease severity and the effect of treatments.

Considerable effort has been made to better understand why some people suffer from severe COVID-19 while others remain asymptomatic. This has led to important clinical findings; people with severe COVID-19 generally experience persistently high levels of inflammation, slower viral load decay, display a dysregulated type-I interferon response, have less active natural killer cells and increased levels of neutrophil extracellular traps. How these findings are connected to the pathogenesis of COVID-19 remains unclear. We propose a mathematical model that sheds light on this issue by focusing on cells that trigger inflammation through molecular patterns: infected cells carrying pathogen-associated molecular patterns (PAMPs) and damaged cells producing damage-associated molecular patterns (DAMPs). The former signals the presence of pathogens while the latter signals danger such as hypoxia or lack of nutrients. Analyses show that SARS-CoV-2 infections can lead to a self-perpetuating feedback loop between DAMP expressing cells and inflammation, identifying the inability to quickly clear PAMPs and DAMPs as the main contributor to hyperinflammation. The model explains clinical findings and reveal conditions that can increase the likelihood of desired clinical outcome from treatment administration. In particular, the analysis suggest that antivirals need to be administered early during infection to have an impact on disease severity. The simplicity of the model and its high level of consistency with clinical findings motivate its use for the formulation of new treatment strategies.

Longitudinal Analysis of SARS-CoV-2 Vaccine Breakthrough Infections Reveals Limited Infectious Virus Shedding and Restricted Tissue Distribution.

Background: The global effort to vaccinate people against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) during an ongoing pandemic has raised questions about how vaccine breakthrough infections compare with infections in immunologically naive individuals and the potential for vaccinated individuals to transmit the virus. Methods: We examined viral dynamics and infectious virus shedding through daily longitudinal sampling in 23 adults infected with SARS-CoV-2 at varying stages of vaccination, including 6 fully vaccinated individuals. Results: The durations of both infectious virus shedding and symptoms were significantly reduced in vaccinated individuals compared with unvaccinated individuals. We also observed that breakthrough infections are associated with strong tissue compartmentalization and are only detectable in saliva in some cases. Conclusions: Vaccination shortens the duration of time of high transmission potential, minimizes symptom duration, and may restrict tissue dissemination.

Multi-species outbreak of SARS-CoV-2 Delta variant in a zoological institution, with the detection in two new families of carnivores.

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has a worldwide distribution in humans and many other mammalian species. In late September 2021, 12 animals maintained by the Chicago Zoological Society's Brookfield Zoo were observed with variable clinical signs. The Delta variant of SARS-CoV-2 was detected in faeces and nasal swabs by qRT-PCR, including the first detection in animals from the families Procyonidae and Viverridae. Test positivity rate was 12.5% for 35 animals tested. All animals had been vaccinated with at least one dose of a recombinant vaccine designed for animals and all recovered with variable supportive treatment. Sequence analysis showed that six zoo animal strains were closely correlated with 18 human SARS-CoV-2 strains, suggestive of potential human-to-animal transmission events. This report documents the expanding host range of COVID-19 during the ongoing pandemic.

Global estimates of the fitness advantage of SARS-CoV-2 variant Omicron.

New variants of SARS-CoV-2 show remarkable heterogeneity in their relative fitness both over time and space. In this paper we extend a previously published model for estimating the selection strength for new SARS-CoV-2 variants to a hierarchical, mixed-effects, renewal equation model. This formulation allows us to globally estimate selection effects at different spatial levels while controlling for complex patterns of transmission and jointly inferring the effects of unit-level covariates in the spatial heterogeneity of SARS-CoV-2 selection effects. Applying this model to the spread of Omicron in 40 counties finding evidence for very strong (64%) but very heterogeneous selection effects at the country level. We further considered different measures of vaccination levels and measures of recent population-level infection as possible explanations. However, none of those variables were found to explain a significant proportion of the heterogeneity in country-level selection effects. We did find a significant positive correlation between the selection advantage of Delta and Omicron at the country level, suggesting that region-specific explanatory variables of fitness differences do exist. Our method is implemented in the Stan programming language, can be run on standard commercial-grade computing resources, and should be straightforward to apply to future variants.