Variant-driven early warning via unsupervised machine learning analysis of spike protein mutations for COVID-19

Paying attention to the SARS-CoV-2 dialect : a deep neural network approach to predicting novel protein mutations

Predicting novel mutations has long-lasting impacts on life science research. Traditionally, this problem is addressed through wet-lab experiments, which are often expensive and time consuming. The recent advancement in neural language models has provided stunning results in modeling and deciphering sequences. In this paper, we propose a Deep Novel Mutation Search (DNMS) method, using deep neural networks, to model protein sequence for mutation prediction. We use SARS-CoV-2 spike protein as the target and use a protein language model to predict novel mutations. Different from existing research which is often limited to mutating the reference sequence for prediction, we propose a parent-child mutation prediction paradigm where a parent sequence is modeled for mutation prediction. Because mutations introduce changing context to the underlying sequence, DNMS models three aspects of the protein sequences: semantic changes, grammatical changes, and attention changes, each modeling protein sequence aspects from shifting of semantics, grammar coherence, and amino-acid interactions in latent space. A ranking approach is proposed to combine all three aspects to capture mutations demonstrating evolving traits, in accordance with real-world SARS-CoV-2 spike protein sequence evolution. DNMS can be adopted for an early warning variant detection system, creating public health awareness of future SARS-CoV-2 mutations.

Using intrahost single nucleotide variant data to predict SARS-CoV-2 detection cycle threshold values

Over the last four years, each successive wave of the COVID-19 pandemic has been caused by variants with mutations that improve the transmissibility of the virus. Despite this, we still lack tools for predicting clinically important features of the virus. In this study, we show that it is possible to predict the PCR cycle threshold (Ct) values from clinical detection assays using sequence data. Ct values often correspond with patient viral load and the epidemiological trajectory of the pandemic. Using a collection of 36,335 high quality genomes, we built models from SARS-CoV-2 intrahost single nucleotide variant (iSNV) data, computing XGBoost models from the frequencies of A, T, G, C, insertions, and deletions at each position relative to the Wuhan-Hu-1 reference genome. Our best model had an R2 of 0.604 [0.593-0.616, 95% confidence interval] and a Root Mean Square Error (RMSE) of 5.247 [5.156-5.337], demonstrating modest predictive power. Overall, we show that the results are stable relative to an external holdout set of genomes selected from SRA and are robust to patient status and the detection instruments that were used. This study highlights the importance of developing modeling strategies that can be applied to publicly available genome sequence data for use in disease prevention and control.

From chaos to clarity: The scientometric breakthrough in COVID-19 research

BACKGROUND

The COVID-19 pandemic paralyzed the world for over three years, generating unprecedented social changes in recent human history.

AIMS

We aimed to scientometrically summarize a global and temporal overview of publications on COVID-19 in the two worst years of the pandemic and its progression in early 2022, after the start of vaccination.

METHODS

Using the Web of Science database, this review covered the period from late 2019 to March 2022 and included all publications identified using the following terms: "SARS-CoV-2", "COVID-19", "Coronavirus Disease 19", and "2019-nCoV". We retrieved 268,904 publications, with evident global spreading, demonstrating that the pandemic triggered worldwide scientific research efforts.

RESULTS

Within the dataset, 195 countries have published about Covid-19. In initial publications, a solid trend in genotyping, sequencing, and detection of the virus was evident; however, in the development of the pandemic, new knowledge and research focus gained relevance, with continental solid trends, revealed by the keywords sustainability (eastern Europe); material sciences (Asia); public and mental health (Africa); information sciences (western Europe); education (Latin America). It identified high-impact research, mainly on diagnosis and vaccines, but also equally essential topics for returning life to the new normal, such as mental health, education, and remote work. The world experienced a highly transmissible infection that proved how fragile we are regarding organization and society.

CONCLUSIONS

It is necessary to learn from such an event and establish a protocol of actions and measures to be taken and avoided in a health emergency, aiming to act differently from the chaos experienced during the pandemic. Following the One Health approach, humanity must be aware of the need for more sustainable attitudes, given the inseparability of human beings from the environment.

Redefining pandemic preparedness: Multidisciplinary insights from the CERP modelling workshop in infectious diseases, workshop report

In July 2023, the Center of Excellence in Respiratory Pathogens organized a two-day workshop on infectious diseases modelling and the lessons learnt from the Covid-19 pandemic. This report summarizes the rich discussions that occurred during the workshop. The workshop participants discussed multisource data integration and highlighted the benefits of combining traditional surveillance with more novel data sources like mobility data, social media, and wastewater monitoring. Significant advancements were noted in the development of predictive models, with examples from various countries showcasing the use of machine learning and artificial intelligence in detecting and monitoring disease trends. The role of open collaboration between various stakeholders in modelling was stressed, advocating for the continuation of such partnerships beyond the pandemic. A major gap identified was the absence of a common international framework for data sharing, which is crucial for global pandemic preparedness. Overall, the workshop underscored the need for robust, adaptable modelling frameworks and the integration of different data sources and collaboration across sectors, as key elements in enhancing future pandemic response and preparedness.

Data-driven recombination detection in viral genomes

Recombination is a key molecular mechanism for the evolution and adaptation of viruses. The first recombinant SARS-CoV-2 genomes were recognized in 2021; as of today, more than ninety SARS-CoV-2 lineages are designated as recombinant. In the wake of the COVID-19 pandemic, several methods for detecting recombination in SARS-CoV-2 have been proposed; however, none could faithfully confirm manual analyses by experts in the field. We hereby present RecombinHunt, an original data-driven method for the identification of recombinant genomes, capable of recognizing recombinant SARS-CoV-2 genomes (or lineages) with one or two breakpoints with high accuracy and within reduced turn-around times. ReconbinHunt shows high specificity and sensitivity, compares favorably with other state-of-the-art methods, and faithfully confirms manual analyses by experts. RecombinHunt identifies recombinant viral genomes from the recent monkeypox epidemic in high concordance with manually curated analyses by experts, suggesting that our approach is robust and can be applied to any epidemic/pandemic virus.

VariantHunter: a method and tool for fast detection of emerging SARS-CoV-2 variants

With the progression of the COVID-19 pandemic, large datasets of SARS-CoV-2 genome sequences were collected to closely monitor the evolution of the virus and identify the novel variants/strains. By analyzing genome sequencing data, health authorities can 'hunt' novel emerging variants of SARS-CoV-2 as early as possible, and then monitor their evolution and spread. We designed VariantHunter, a highly flexible and user-friendly tool for systematically monitoring the evolution of SARS-CoV-2 at global and regional levels. In VariantHunter, amino acid changes are analyzed over an interval of 4 weeks in an arbitrary geographical area (continent, country, or region); for every week in the interval, the prevalence is computed and changes are ranked based on their increase or decrease in prevalence. VariantHunter supports two main types of analysis: lineage-independent and lineage-specific. The former considers all the available data and aims to discover new viral variants. The latter evaluates specific lineages/viral variants to identify novel candidate designations (sub-lineages and sub-variants). Both analyses use simple statistics and visual representations (diffusion charts and heatmaps) to track viral evolution. A dataset explorer allows users to visualize available data and refine their selection. VariantHunter is a web application free to all users. The two types of supported analysis (lineage-independent and lineage-specific) allow user-friendly monitoring of the viral evolution, empowering genomic surveillance without requiring any computational background. Database URL http://gmql.eu/variant_hunter/.

Using Haplotype-Based Artificial Intelligence to Evaluate SARS-CoV-2 Novel Variants and Mutations

Importance

Earlier detection of emerging novel SARS-COV-2 variants is important for public health surveillance of potential viral threats and for earlier prevention research. Artificial intelligence may facilitate early detection of SARS-CoV2 emerging novel variants based on variant-specific mutation haplotypes and, in turn, be associated with enhanced implementation of risk-stratified public health prevention strategies.

Objective

To develop a haplotype-based artificial intelligence (HAI) model for identifying novel variants, including mixture variants (MVs) of known variants and new variants with novel mutations.

Design, Setting, and Participants

This cross-sectional study used serially observed viral genomic sequences globally (prior to March 14, 2022) to train and validate the HAI model and used it to identify variants arising from a prospective set of viruses from March 15 to May 18, 2022.

Main Outcomes and Measures

Viral sequences, collection dates, and locations were subjected to statistical learning analysis to estimate variant-specific core mutations and haplotype frequencies, which were then used to construct an HAI model to identify novel variants.

Results

Through training on more than 5 million viral sequences, an HAI model was built, and its identification performance was validated on an independent validation set of more than 5 million viruses. Its identification performance was assessed on a prospective set of 344 901 viruses. In addition to achieving an accuracy of 92.8% (95% CI within 0.1%), the HAI model identified 4 Omicron MVs (Omicron-Alpha, Omicron-Delta, Omicron-Epsilon, and Omicron-Zeta), 2 Delta MVs (Delta-Kappa and Delta-Zeta), and 1 Alpha-Epsilon MV, among which Omicron-Epsilon MVs were most frequent (609/657 MVs [92.7%]). Furthermore, the HAI model found that 1699 Omicron viruses had unidentifiable variants given that these variants acquired novel mutations. Lastly, 524 variant-unassigned and variant-unidentifiable viruses carried 16 novel mutations, 8 of which were increasing in prevalence percentages as of May 2022.

Conclusions and Relevance

In this cross-sectional study, an HAI model found SARS-COV-2 viruses with MV or novel mutations in the global population, which may require closer examination and monitoring. These results suggest that HAI may complement phylogenic variant assignment, providing additional insights into emerging novel variants in the population.

Early detection of variants of concern via funnel plots of regional reproduction numbers

Early detection of the emergence of a new variant of concern (VoC) is essential to develop strategies that contain epidemic outbreaks. For example, knowing in which region a VoC starts spreading enables prompt actions to circumscribe the geographical area where the new variant can spread, by containing it locally. This paper presents 'funnel plots' as a statistical process control method that, unlike tools whose purpose is to identify rises of the reproduction number ([Formula: see text]), detects when a regional [Formula: see text] departs from the national average and thus represents an anomaly. The name of the method refers to the funnel-like shape of the scatter plot that the data take on. Control limits with prescribed false alarm rate are derived from the observation that regional [Formula: see text]'s are normally distributed with variance inversely proportional to the number of infectious cases. The method is validated on public COVID-19 data demonstrating its efficacy in the early detection of SARS-CoV-2 variants in India, South Africa, England, and Italy, as well as of a malfunctioning episode of the diagnostic infrastructure in England, during which the Immensa lab in Wolverhampton gave 43,000 incorrect negative tests relative to South West and West Midlands territories.

SARS-CoV-2 Omicron Variant Genomic Sequences and Their Epidemiological Correlates Regarding the End of the Pandemic: In Silico Analysis

Background

Emergence of the new SARS-CoV-2 variant B.1.1.529 worried health policy makers worldwide due to a large number of mutations in its genomic sequence, especially in the spike protein region. The World Health Organization (WHO) designated this variant as a global variant of concern (VOC), which was named "Omicron." Following Omicron's emergence, a surge of new COVID-19 cases was reported globally, primarily in South Africa.

Objective

The aim of this study was to understand whether Omicron had an epidemiological advantage over existing variants.

Methods

We performed an in silico analysis of the complete genomic sequences of Omicron available on the Global Initiative on Sharing Avian Influenza Data (GISAID) database to analyze the functional impact of the mutations present in this variant on virus-host interactions in terms of viral transmissibility, virulence/lethality, and immune escape. In addition, we performed a correlation analysis of the relative proportion of the genomic sequences of specific SARS-CoV-2 variants (in the period from October 1 to November 29, 2021) with matched epidemiological data (new COVID-19 cases and deaths) from South Africa.

Results

Compared with the current list of global VOCs/variants of interest (VOIs), as per the WHO, Omicron bears more sequence variation, specifically in the spike protein and host receptor-binding motif (RBM). Omicron showed the closest nucleotide and protein sequence homology with the Alpha variant for the complete sequence and the RBM. The mutations were found to be primarily condensed in the spike region (n=28-48) of the virus. Further mutational analysis showed enrichment for the mutations decreasing binding affinity to angiotensin-converting enzyme 2 receptor and receptor-binding domain protein expression, and for increasing the propensity of immune escape. An inverse correlation of Omicron with the Delta variant was noted (r=-0.99, P<.001; 95% CI -0.99 to -0.97) in the sequences reported from South Africa postemergence of the new variant, subsequently showing a decrease. There was a steep rise in new COVID-19 cases in parallel with the increase in the proportion of Omicron isolates since the report of the first case (74%-100%). By contrast, the incidence of new deaths did not increase (r=-0.04, P>.05; 95% CI -0.52 to 0.58).

Conclusions

In silico analysis of viral genomic sequences suggests that the Omicron variant has more remarkable immune-escape ability than existing VOCs/VOIs, including Delta, but reduced virulence/lethality than other reported variants. The higher power for immune escape for Omicron was a likely reason for the resurgence in COVID-19 cases and its rapid rise as the globally dominant strain. Being more infectious but less lethal than the existing variants, Omicron could have plausibly led to widespread unnoticed new, repeated, and vaccine breakthrough infections, raising the population-level immunity barrier against the emergence of new lethal variants. The Omicron variant could have thus paved the way for the end of the pandemic.

Genomic Epidemiology of the SARS-CoV-2 Epidemic in Cyprus from November 2020 to October 2021: The Passage of Waves of Alpha and Delta Variants of Concern

The emergence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in December 2019 resulted in the coronavirus disease 2019 (COVID-19) pandemic, which has had devastating repercussions for public health. Over the course of this pandemic, the virus has continuously been evolving, resulting in new, more infectious variants that have frequently led to surges of new SARS-CoV-2 infections. In the present study, we performed detailed genetic, phylogenetic, phylodynamic and phylogeographic analyses to examine the SARS-CoV-2 epidemic in Cyprus using 2352 SARS-CoV-2 sequences from infected individuals in Cyprus during November 2020 to October 2021. During this period, a total of 61 different lineages and sublineages were identified, with most falling into three groups: B.1.258 & sublineages, Alpha (B.1.1.7 & Q. sublineages), and Delta (B.1.617.2 & AY. sublineages), each encompassing a set of S gene mutations that primarily confer increased transmissibility as well as immune evasion. Specifically, these lineages were coupled with surges of new infections in Cyprus, resulting in the following: the second wave of SARS-CoV-2 infections in Cyprus, comprising B.1.258 & sublineages, during late autumn 2020/beginning of winter 2021; the third wave, comprising Alpha (B.1.1.7 & Q. sublineages), during spring 2021; and the fourth wave, comprising Delta (B.1.617.2 & AY. sublineages) during summer 2021. Additionally, it was identified that these lineages were primarily imported from and exported to the UK, Greece, and Sweden; many other migration links were also identified, including Switzerland, Denmark, Russia, and Germany. Taken together, the results of this study indicate that the SARS-CoV-2 epidemic in Cyprus was characterized by successive introduction of new lineages from a plethora of countries, resulting in the generation of waves of infection. Overall, this study highlights the importance of investigating the spatiotemporal evolution of the SARS-CoV-2 epidemic in the context of Cyprus, as well as the impact of protective measures placed to mitigate transmission of the virus, providing necessary information to safeguard public health.

Investigating SARS-CoV-2 breakthrough infections per variant and vaccine type

Breakthrough SARS-CoV-2 infections have been reported in fully vaccinated individuals, in spite of the high efficacy of the currently available vaccines, proven in trials and real-world studies. Several variants of concern (VOC) have been proffered to be associated with breakthrough infections following immunization. In this study, we investigated 378 breakthrough infections recorded between January and July 2021 and compared the distribution of SARS-CoV-2 genotypes identified in 225 fully vaccinated individuals to the frequency of circulating community lineages in the region of South Limburg (The Netherlands) in a week-by-week comparison. Although the proportion of breakthrough infections was relatively low and stable when the Alpha variant was predominant, the rapid emergence of the Delta variant lead to a strong increase in breakthrough infections, with a higher relative proportion of individuals vaccinated with Vaxzevria or Jcovden being infected compared to those immunized with mRNA-based vaccines. A significant difference in median age was observed when comparing fully vaccinated individuals with severe symptoms (83 years) to asymptomatic cases (46.5 years) or individuals with mild-to-moderate symptoms (42 years). There was no association between SARS-CoV-2 genotype or vaccine type and disease symptoms. Furthermore, the majority of adaptive mutations were concentrated in the N-terminal domain of the Spike protein, highlighting its role in immune evasion. Interestingly, symptomatic individuals harbored significantly higher SARS-CoV-2 loads than asymptomatic vaccinated individuals and breakthrough infections caused by the Delta variant were associated with increased viral loads compared to those caused by the Alpha variant. In addition, we investigated the role of the Omicron variant in causing breakthrough infections by analyzing 135 samples that were randomly selected for genomic surveillance during the transition period from Delta to Omicron. We found that the proportion of Omicron vs. Delta infections was significantly higher in individuals who received a booster vaccine compared to both unvaccinated and fully vaccinated individuals. Altogether, these results indicate that the emergence of the Delta variant and in particular Omicron has lowered the efficiency of particular vaccine types to prevent SARS-CoV-2 infections and that, although rare, the elderly are particularly at risk of becoming severely infected as the consequence of a breakthrough infection.

Analysis of co-occurring and mutually exclusive amino acid changes and detection of convergent and divergent evolution events in SARS-CoV-2

The inflation of SARS-CoV-2 lineages with a high number of accumulated mutations (such as the recent case of Omicron) has risen concerns about the evolutionary capacity of this virus. Here, we propose a computational study to examine non-synonymous mutations gathered within genomes of SARS-CoV-2 from the beginning of the pandemic until February 2022. We provide both qualitative and quantitative descriptions of such corpus, focusing on statistically significant co-occurring and mutually exclusive mutations within single genomes. Then, we examine in depth the distributions of mutations over defined lineages and compare those of frequently co-occurring mutation pairs. Based on this comparison, we study mutations' convergence/divergence on the phylogenetic tree. As a result, we identify 1,818 co-occurring pairs of non-synonymous mutations showing at least one event of convergent evolution and 6,625 co-occurring pairs with at least one event of divergent evolution. Notable examples of both types are shown by means of a tree-based representation of lineages, visually capturing mutations' behaviors. Our method confirms several well-known cases; moreover, the provided evidence suggests that our workflow can explain aspects of the future mutational evolution of SARS-CoV-2.

Epidemiological theory of virus variants

We propose a physics-inspired mathematical model underlying the temporal evolution of competing virus variants that relies on the existence of (quasi) fixed points capturing the large time scale invariance of the dynamics. To motivate our result we first modify the time-honoured compartmental models of the SIR type to account for the existence of competing variants and then show how their evolution can be naturally re-phrased in terms of flow equations ending at quasi fixed points. As the natural next step we employ (near) scale invariance to organise the time evolution of the competing variants within the effective description of the epidemic Renormalisation Group framework. We test the resulting theory against the time evolution of COVID-19 virus variants that validate the theory empirically.

A Bioinformatics Tool for Predicting Future COVID-19 Waves Based on a Retrospective Analysis of the Second Wave in India: Model Development Study

Background

Since the start of the COVID-19 pandemic, health policymakers globally have been attempting to predict an impending wave of COVID-19. India experienced a devastating second wave of COVID-19 in the late first week of May 2021. We retrospectively analyzed the viral genomic sequences and epidemiological data reflecting the emergence and spread of the second wave of COVID-19 in India to construct a prediction model.

Objective

We aimed to develop a bioinformatics tool that can predict an impending COVID-19 wave.

Methods

We analyzed the time series distribution of genomic sequence data for SARS-CoV-2 and correlated it with epidemiological data for new cases and deaths for the corresponding period of the second wave. In addition, we analyzed the phylodynamics of circulating SARS-CoV-2 variants in the Indian population during the study period.

Results

Our prediction analysis showed that the first signs of the arrival of the second wave could be seen by the end of January 2021, about 2 months before its peak in May 2021. By the end of March 2021, it was distinct. B.1.617 lineage variants powered the wave, most notably B.1.617.2 (Delta variant).

Conclusions

Based on the observations of this study, we propose that genomic surveillance of SARS-CoV-2 variants, complemented with epidemiological data, can be a promising tool to predict impending COVID-19 waves.