A Deep Learning Approach to Forecast Short-Term COVID-19 Cases and Deaths in the US

Since the US reported its first COVID-19 case on January 21, 2020, the science community has been applying various techniques to forecast incident cases and deaths. To date, providing an accurate and robust forecast at a high spatial resolution has proved challenging, even in the short term. Here we present a novel multi-stage deep learning model to forecast the number of COVID-19 cases and deaths for each US state at a weekly level for a forecast horizon of 1 to 4 weeks. The model is heavily data driven, and relies on epidemiological, mobility, survey, climate, and demographic. We further present results from a case study that incorporates SARS-CoV-2 genomic data (i.e. variant cases) to demonstrate the value of incorporating variant cases data into model forecast tools. We implement a rigorous and robust evaluation of our model - specifically we report on weekly performance over a one-year period based on multiple error metrics, and explicitly assess how our model performance varies over space, chronological time, and different outbreak phases. The proposed model is shown to consistently outperform the CDC ensemble model for all evaluation metrics in multiple spatiotemporal settings, especially for the longer-term (3 and 4 weeks ahead) forecast horizon. Our case study also highlights the potential value of virus genomic data for use in short-term forecasting to identify forthcoming surges driven by new variants. Based on our findings, the proposed forecasting framework improves upon the available forecasting tools currently used to support public health decision making with respect to COVID-19 risk. Research in contextO_ST_ABSEvidence before this studyC_ST_ABSA systematic review of the COVID-19 forecasting and the EPIFORGE 2020 guidelines reveal the lack of consistency, reproducibility, comparability, and quality in the current COVID-19 forecasting literature. To provide an updated survey of the literature, we carried out our literature search on Google Scholar, PubMed, and medRxi, using the terms "Covid-19," "SARS-CoV-2," "coronavirus," "short-term," "forecasting," and "genomic surveillance." Although the literature includes a significant number of papers, it remains lacking with respect to rigorous model evaluation, interpretability and translation. Furthermore, while SARS-CoV-2 genomic surveillance is emerging as a vital necessity to fight COVID-19 (i.e. wastewater sampling and airport screening), to our knowledge, no published forecasting model has illustrated the value of virus genomic data for informing future outbreaks. Added value of this studyWe propose a multi-stage deep learning model to forecast COVID-19 cases and deaths with a horizon window of four weeks. The data driven model relies on a comprehensive set of input features, including epidemiological, mobility, behavioral survey, climate, and demographic. We present a robust evaluation framework to systematically assess the model performance over a one-year time span, and using multiple error metrics. This rigorous evaluation framework reveals how the predictive accuracy varies over chronological time, space, and outbreak phase. Further, a comparative analysis against the CDC ensemble, the best performing model in the COVID-19 ForecastHub, shows the model to consistently outperform the CDC ensemble for all evaluation metrics in multiple spatiotemporal settings, especially for the longer forecasting windows. We also conduct a feature analysis, and show that the role of explanatory features changes over time. Specifically, we note a changing role of climate variables on model performance in the latter half of the study period. Lastly, we present a case study that reveals how incorporating SARS-CoV-2 genomic surveillance data may improve forecasting accuracy compared to a model without variant cases data. Implications of all the available evidenceResults from the robust evaluation analysis highlight extreme model performance variability over time and space, and suggest that forecasting models should be accompanied with specifications on the conditions under which they perform best (and worst), in order to maximize their value and utility in aiding public health decision making. The feature analysis reveals the complex and changing role of factors contributing to COVID-19 transmission over time, and suggests a possible seasonality effect of climate on COVID-19 spread, but only after August 2021. Finally, the case study highlights the added value of using genomic surveillance data in short-term epidemiological forecasting models, especially during the early stage of new variant introductions.

Insights into the limited global spread of the immune evasive SARS-CoV-2 variant Mu

SARS-CoV-2 Variants of Concern (VOCs) continue to reshape the trajectory of the COVID-19 pandemic. However, why some VOCs, like Omicron, become globally dominant while the spread of others is limited is not fully understood. To address this question, we investigated the VOC Mu, which was first identified in Colombia in late 2020. Our study demonstrates that, although Mu is less sensitive to neutralization compared to variants that preceded it, it did not spread significantly outside of South and Central America. Additionally, we find evidence that the response to Mu was impeded by reporting delays and gaps in the global genomic surveillance system. Our findings suggest that immune evasion alone was not sufficient to outcompete highly transmissible variants that were circulating concurrently with Mu. Insights into the complex relationship between genomic and epidemiological characteristics of previous variants should inform our response to variants that are likely to emerge in the future.

Global disparities in SARS-CoV-2 genomic surveillance

Genomic sequencing provides critical information to track the evolution and spread of SARS-CoV-2, optimize molecular tests, treatments and vaccines, and guide public health responses. To investigate the spatiotemporal heterogeneity in the global SARS-CoV-2 genomic surveillance, we estimated the impact of sequencing intensity and turnaround times (TAT) on variant detection in 167 countries. Most countries submit genomes >21 days after sample collection, and 77% of low and middle income countries sequenced <0.5% of their cases. We found that sequencing at least 0.5% of the cases, with a TAT <21 days, could be a benchmark for SARS-CoV-2 genomic surveillance efforts. Socioeconomic inequalities substantially impact our ability to quickly detect SARS-CoV-2 variants, and undermine the global pandemic preparedness. One-Sentence SummarySocioeconomic inequalities impacted the SARS-CoV-2 genomic surveillance, and undermined the global pandemic preparedness.

Sequencing SARS-CoV-2 Genomes from Saliva

Genomic sequencing is crucial to understanding the epidemiology and evolution of SARS-CoV-2. Often, genomic studies rely on remnant diagnostic material, typically nasopharyngeal swabs, as input into whole genome SARS-CoV-2 next-generation sequencing pipelines. Saliva has proven to be a safe and stable specimen for the detection of SARS-CoV-2 RNA via traditional diagnostic assays, however saliva is not commonly used for SARS-CoV-2 sequencing. Using the ARTIC Network amplicon-generation approach with sequencing on the Oxford Nanopore MinION, we demonstrate that sequencing SARS-CoV-2 from saliva produces genomes comparable to those from nasopharyngeal swabs, and that RNA extraction is necessary to generate complete genomes from saliva. In this study, we show that saliva is a useful specimen type for genomic studies of SARS-CoV-2.

Emergence of an early SARS-CoV-2 epidemic in the United States

The emergence of the early COVID-19 epidemic in the United States (U.S.) went largely undetected, due to a lack of adequate testing and mitigation efforts. The city of New Orleans, Louisiana experienced one of the earliest and fastest accelerating outbreaks, coinciding with the annual Mardi Gras festival, which went ahead without precautions. To gain insight into the emergence of SARS-CoV-2 in the U.S. and how large, crowded events may have accelerated early transmission, we sequenced SARS-CoV-2 genomes during the first wave of the COVID-19 epidemic in Louisiana. We show that SARS-CoV-2 in Louisiana initially had limited sequence diversity compared to other U.S. states, and that one successful introduction of SARS-CoV-2 led to almost all of the early SARS-CoV-2 transmission in Louisiana. By analyzing mobility and genomic data, we show that SARS-CoV-2 was already present in New Orleans before Mardi Gras and that the festival dramatically accelerated transmission, eventually leading to secondary localized COVID-19 epidemics throughout the Southern U.S.. Our study provides an understanding of how superspreading during large-scale events played a key role during the early outbreak in the U.S. and can greatly accelerate COVID-19 epidemics on a local and regional scale.

Kinetics of antibody responses dictate COVID-19 outcome

Recent studies have provided insights into innate and adaptive immune dynamics in coronavirus disease 2019 (COVID-19). Yet, the exact feature of antibody responses that governs COVID-19 disease outcomes remain unclear. Here, we analysed humoral immune responses in 209 asymptomatic, mild, moderate and severe COVID-19 patients over time to probe the nature of antibody responses in disease severity and mortality. We observed a correlation between anti-Spike (S) IgG levels, length of hospitalization and clinical parameters associated with worse clinical progression. While high anti-S IgG levels correlated with worse disease severity, such correlation was time-dependent. Deceased patients did not have higher overall humoral response than live discharged patients. However, they mounted a robust, yet delayed response, measured by anti-S, anti-RBD IgG, and neutralizing antibody (NAb) levels, compared to survivors. Delayed seroconversion kinetics correlated with impaired viral control in deceased patients. Finally, while sera from 89% of patients displayed some neutralization capacity during their disease course, NAb generation prior to 14 days of disease onset emerged as a key factor for recovery. These data indicate that COVID-19 mortality does not correlate with the cross-sectional antiviral antibody levels per se, but rather with the delayed kinetics of NAb production.

SalivaDirect: Simple and sensitive molecular diagnostic test for SARS-CoV-2 surveillance

Current bottlenecks for improving accessibility and scalability of SARS-CoV-2 testing include diagnostic assay costs, complexity, and supply chain shortages. To resolve these issues, we developed SalivaDirect, which received Emergency Use Authorization (EUA) from the U.S. Food and Drug Administration on August 15th, 2020. The critical component of our approach is to use saliva instead of respiratory swabs, which enables non-invasive frequent sampling and reduces the need for trained healthcare professionals during collection. Furthermore, we simplified our diagnostic test by (1) not requiring nucleic acid preservatives at sample collection, (2) replacing nucleic acid extraction with a simple proteinase K and heat treatment step, and (3) testing specimens with a dualplex quantitative reverse transcription PCR (RT-qPCR) assay. We validated SalivaDirect with reagents and instruments from multiple vendors to minimize the risk for supply chain issues. Regardless of our tested combination of reagents and instruments from different vendors, we found that SalivaDirect is highly sensitive with a limit of detection of 6-12 SARS-CoV-2 copies/L. When comparing SalivaDirect to paired nasopharyngeal swabs using the authorized ThermoFisher Scientific TaqPath COVID-19 combo kit, we found high agreement in testing outcomes (>94%). In partnership with the National Basketball Association (NBA) and Players Association, we conducted a large-scale (n = 3,779) SalivaDirect usability study and comparison to standard nasal/oral tests for asymptomatic and presymptomatic SARS-CoV-2 detection. From this cohort of healthy NBA players, staff, and contractors, we found that 99.7% of samples were valid using our saliva collection techniques and a 89.5% positive and >99.9% negative test agreement to swabs, demonstrating that saliva is a valid and noninvasive alternative to swabs for large-scale SARS-CoV-2 testing. SalivaDirect is a flexible and inexpensive ($1.21-$4.39/sample in reagent costs) option to help improve SARS-CoV-2 testing capacity. Register to become a designated laboratory to use SalivaDirect under our FDA EUA on our website: publichealth.yale.edu/salivadirect/.

Longitudinal immunological analyses reveal inflammatory misfiring in severe COVID-19 patients

Recent studies have provided insights into the pathogenesis of coronavirus disease 2019 (COVID-19)1-4. Yet, longitudinal immunological correlates of disease outcome remain unclear. Here, we serially analysed immune responses in 113 COVID-19 patients with moderate (non-ICU) and severe (ICU) disease. Immune profiling revealed an overall increase in innate cell lineages with a concomitant reduction in T cell number. We identify an association between early, elevated cytokines and worse disease outcomes. Following an early increase in cytokines, COVID-19 patients with moderate disease displayed a progressive reduction in type-1 (antiviral) and type-3 (antifungal) responses. In contrast, patients with severe disease maintained these elevated responses throughout the course of disease. Moreover, severe disease was accompanied by an increase in multiple type 2 (anti-helminths) effectors including, IL-5, IL-13, IgE and eosinophils. Unsupervised clustering analysis of plasma and peripheral blood leukocyte data identified 4 immune signatures, representing (A) growth factors, (B) type-2/3 cytokines, (C) mixed type-1/2/3 cytokines, and (D) chemokines that correlated with three distinct disease trajectories of patients. The immune profile of patients who recovered with moderate disease was enriched in tissue reparative growth factor signature (A), while the profile for those with worsened disease trajectory had elevated levels of all four signatures. Thus, we identified development of a maladapted immune response profile associated with severe COVID-19 outcome and early immune signatures that correlate with divergent disease trajectories.

Sex differences in immune responses to SARS-CoV-2 that underlie disease outcomes

A growing body of evidence indicates sex differences in the clinical outcomes of coronavirus disease 2019 (COVID-19)1-4. However, whether immune responses against SARS-CoV-2 differ between sexes, and whether such differences explain male susceptibility to COVID-19, is currently unknown. In this study, we examined sex differences in viral loads, SARS-CoV-2-specific antibody titers, plasma cytokines, as well as blood cell phenotyping in COVID-19 patients. By focusing our analysis on patients with mild to moderate disease who had not received immunomodulatory medications, our results revealed that male patients had higher plasma levels of innate immune cytokines and chemokines including IL-8, IL-18, and CCL5, along with more robust induction of non-classical monocytes. In contrast, female patients mounted significantly more robust T cell activation than male patients during SARS-CoV-2 infection, which was sustained in old age. Importantly, we found that a poor T cell response negatively correlated with patients age and was predictive of worse disease outcome in male patients, but not in female patients. Conversely, higher innate immune cytokines in female patients associated with worse disease progression, but not in male patients. These findings reveal a possible explanation underlying observed sex biases in COVID-19, and provide important basis for the development of sex-based approach to the treatment and care of men and women with COVID-19.

Detection of SARS-CoV-2 RNA by multiplex RT-qPCR

The current RT-qPCR assay recommended for SARS-CoV-2 testing in the United States requires analysis of three genomic targets per sample: two viral and one host. To simplify testing and reduce the volume of required reagents, we developed a multiplex RT-qPCR assay to detect SARS-CoV-2 in a single reaction. We used existing N1, N2, and RP primer and probe sets by the CDC, but substituted fluorophores to allow multiplexing of the assay. The cycle threshold (Ct) values of our multiplex RT-qPCR were comparable to those obtained by the singleplex assay adapted for research purposes. Low copies (>500 copies / reaction) of SARS-CoV-2 RNA were consistently detected by the multiplex RT-qPCR. Our novel multiplex RT-qPCR improves upon current singleplex diagnostics by saving reagents, costs, time and labor.

First case of placental infection with SARS-CoV-2

BackgroundThe effects of Covid-19 in pregnancy remain relatively unknown. We present a case of second trimester pregnancy with symptomatic Covid-19 complicated by severe preeclampsia and placental abruption. MethodsWe analyzed placenta for the presence of SARS-CoV-2 through molecular and immunohistochemical assays and by and electron microscopy, and we measured the maternal antibody response in blood to this infection. ResultsSARS-CoV-2 localized predominantly to syncytiotrophoblast cells at the maternal-fetal interface of the placenta. Histological examination of the placenta revealed a dense macrophage infiltrate, but no evidence for vasculopathy typically associated with preeclampsia. ConclusionThis case demonstrates, for the first time, SARS-CoV-2 invasion of the placenta, highlighting the potential for severe morbidity among pregnant women with Covid-19.