CovidOutcome2: a tool for SARS-CoV2 mutation identification and for disease severity prediction

Our goal was to develop a platform, CovidOutcome2, capable of predicting disease severity from viral mutation profiles using automated machine learning (autoML) and deep neural networks applied to the available large corpus of sequenced SARS-CoV2 genomes. CovidOutcome2 accepts either user-submitted genomes or user defined mutation combinations as the input. The output is a predicted severity score plus a list of identified, annotated mutations and their functional effects in VCF format. The best model performance is a ROC-AUC 0.899 for the model including patient age and ROC-AUC 0.83 for the model without patient age. AvailabilityCovidOutcome is freely available online under the URL https://www.covidoutcome.bio-ml.com as well as in a standalone version https://github.com/bio-apps/covid-outcome.

Microsimulation based quantitative analysis of COVID-19 management strategies

BackgroundPandemic management includes a variety of control measures, such as social distancing, testing/quarantining and vaccination applied to a population where the virus is circulating. The COVID-19 (SARS-CoV-2) pandemic is mitigated by several non-pharmaceutical interventions, but it is hard to predict which of these regulations are the most effective for a given population. MethodsWe developed a computationally effective and scalable, agent-based microsimulation framework. This unified framework was fitted to realistic data to enable us to test control measures (closures, quarantining, testing, vaccination) in multiple infection waves caused by the spread of a new virus variant in a city-sized societal environment. Our framework is capable of simulating nine billion agent-steps per minute, allowing us to model interactions in populations with up to 90 million individuals. FindingsWe show that vaccination strategies prioritising occupational risk groups minimise the number of infections but allow higher mortality while prioritising vulnerable groups minimises mortality but implies increased infection rate. We also found that intensive vaccination and non-pharmaceutical interventions can substantially suppress the spread of the virus, while low levels of vaccination and premature reopening may easily revert the epidemic to an uncontrolled state. InterpretationOur analysis highlights that while vaccination protects the elderly from COVID-19, a large percentage of children will contract and spread the virus, and we also show the benefits and limitations of various quarantine and testing scenarios. FundingThis work was carried out within the framework of the Hungarian National Development, Research, and Innovation (NKFIH) Fund 2020-2.1.1-ED-2020-00003. Research in contextO_ST_ABSEvidence before this studyC_ST_ABSWe still do not have an effective medical treatment against COVID-19 (SARS-CoV-2), thus the majority of the efforts to stop the pandemic focuses on non-pharmaceutical interventions. Each country came up with a local solution to stop the spread of the virus by increased testing, quarantining, lock-down of various events and institutions or early vaccination. There is no clear way how these interventions can be compared, and it is especially challenging to predict how combinations of interventions could influence the pandemic. Various mathematical modelling approaches helped decision-makers to foresee the effects of their decisions. Most of these models rely on classical, deterministic compartmental "SEIR" models, which can be solved easily but cannot take into account spatial effects and most non-pharmaceutical interventions affect the same parameters, so there is no way to analyse their separate or joint effects. Agent-based microsimulations are harder to solve but can consider far more details. Several models were developed on these lines focusing on questions about ideal vaccination, lock-down or other specific problems, but none of these studies evaluated and compared the individual and mixed effects of a wide variety of control measures. Added-value of this studyHere we present PanSim, a framework where we introduce a detailed infection event simulation step and the possibility to control specific workplaces individually (schools, hospitals, etc.), test various vaccination, testing and quarantine scenarios while considering preconditions, age, sex, residence and workplace of individuals and mutant viruses with various infectivity. The level of details and granularity of simulations allow our work to evaluate this wide range of scenarios and control measures accurately and directly compare them with one another to provide quantitative evidence to support decision-makers. Analysis of our simulations also provides emergent results on the risks children and non-vaccinated individuals face. Implications of all the available evidenceThe agent-based microsimulation framework allows us to evaluate the risk and possible consequences of particular interventions precisely. Due to the outstanding efficiency of the computations, it is possible to apply scenario-based analysis and control design methods which require a high number of simulation runs to obtain results on a given confidence level. This will enable us to design and quantitatively assess control measures in case of new waves of COVID-19 or new pandemic outbreaks.

COVIDOUTCOME - Estimating COVID Severity Based on Mutation Signatures in the SARS-CoV-2 Genome

IntroductionNumerous studies demonstrate frequent mutations in the genome of SARS-CoV-2. Our goal was to statistically link mutations to severe disease outcome. MethodsWe used an automated machine learning approach where 1,594 viral genomes with available clinical follow-up data were used as the training set (797 "severe" and 797 "mild"). The best algorithm, based on random forest classification combined with the LASSO feature selection algorithm was employed to the training set to link mutation signatures and outcome. The performance of the final model was estimated by repeated, stratified, 10-fold cross validation (CV), then adjusted for multiple testing with Bootstrap Bias Corrected CV. ResultsWe identified 26 protein and UTR mutations significantly linked to severe outcome. The best classification algorithm uses a mutation signature of 22 mutations as well as the patients age as the input and shows high classification efficiency with an AUC of 0.94 (CI: [0.912, 0.962]) and a prediction accuracy of 87% (CI: [0.830, 0.903]). Finally, we established an online platform (https://covidoutcome.com/) which is capable to use a viral sequence and the patients age as the input and provides a percentage estimation of disease severity. DiscussionWe demonstrate a statistical association between mutation signatures of SARS-CoV-2 and severe outcome of COVID-19. The established analysis platform enables a real-time analysis of new viral genomes. KEY MESSAGESO_LIA statistical link between SARS-Cov-2 mutation status and severe COVID outcome was established using automated machine learning techniques based on random forest and logistic regression combined with feature selection algorithms. C_LIO_LIA mutation signature based on 3,779 protein coding and 36 UTR mutations capable to identify severe outcome cases was established. C_LIO_LIThe trained model showed high classification performance (AUC=0.94 (CI: [0.912, 0.962]), accuracy=0.87 (CI: [0.830, 0.903])). C_LIO_LIA registration-free web-server for automated classification of new samples was set up and is accessible at http://www.covidoutcome.com. C_LIO_LIThe established pipeline provides a quick assessment of future patients warranting a prospective clinical validation. C_LI

Different mutations in SARS-CoV-2 associate with severe and mild outcome

IntroductionGenomic alterations in a viral genome can lead to either better or worse outcome and identifying these mutations is of utmost importance. Here, we correlated protein-level mutations in the SARS-CoV-2 virus to clinical outcome. MethodsMutations in viral sequences from the GISAID virus repository were evaluated by using "hCoV-19/Wuhan/WIV04/2019" as the reference. Patient outcomes were classified as mild disease, hospitalization and severe disease (death or documented treatment in an intensive-care unit). Chi-square test was applied to examine the association between each mutation and patient outcome. False discovery rate was computed to correct for multiple hypothesis testing and results passing a FDR cutoff of 5% were accepted as significant. ResultMutations were mapped to amino acid changes for 2,120 non-silent mutations. Mutations correlated to mild outcome were located in the ORF8, NSP6, ORF3a, NSP4, and in the nucleocapsid phosphoprotein N. Mutations associated with inferior outcome were located in the surface (S) glycoprotein, in the RNA dependent RNA polymerase, in the 3-to5 exonuclease, in ORF3a, NSP2 and N. Mutations leading to severe outcome with low prevalence were found in the surface (S) glycoprotein and in NSP7. Five out of 17 of the most significant mutations mapped onto a 10 amino acid long phosphorylated stretch of N indicating that in spite of obvious sampling restrictions the approach can find functionally relevant sites in the viral genome. ConclusionsWe demonstrate that mutations in the viral genes may have a direct correlation to clinical outcome. Our results help to quickly identify SARS-CoV-2 infections harboring mutations related to severe outcome.