A machine learning-based approach to determine infection status in recipients of BBV152 whole virion inactivated SARS-CoV-2 vaccine for serological surveys

Data science has been an invaluable part of the COVID-19 pandemic response with multiple applications, ranging from tracking viral evolution to understanding the effectiveness of interventions. Asymptomatic breakthrough infections have been a major problem during the ongoing surge of Delta variant globally. Serological discrimination of vaccine response from infection has so far been limited to Spike protein vaccines used in the higher-income regions. Here, we show for the first time how statistical and machine learning (ML) approaches can discriminate SARS-CoV-2 infection from immune response to an inactivated whole virion vaccine (BBV152, Covaxin, India), thereby permitting real-world vaccine effectiveness assessments from cohort-based serosurveys in Asia and Africa where such vaccines are commonly used. Briefly, we accessed serial data on Anti-S and Anti-NC antibody concentration values, along with age, sex, number of doses, and number of days since the last vaccine dose for 1823 Covaxin recipients. An ensemble ML model, incorporating a consensus clustering approach alongside the support vector machine (SVM) model, was built on 1063 samples where reliable qualifying data existed, and then applied to the entire dataset. Of 1448 self-reported negative subjects, 724 were classified as infected. Since the vaccine contains wild-type virus and the antibodies induced will neutralize wild type much better than Delta variant, we determined the relative ability of a random subset of such samples to neutralize Delta versus wild type strain. In 100 of 156 samples, where ML prediction differed from self-reported uninfected status, Delta variant, was neutralized more effectively than the wild type, which cannot happen without infection. The fraction rose to 71.8% (28 of 39) in subjects predicted to be infected during the surge, which is concordant with the percentage of sequences classified as Delta (75.6%-80.2%) over the same period.

Genomic surveillance and phylodynamic analyses reveal emergence of novel mutation and co-mutation patterns within SARS-CoV2 variants prevalent in India

Emergence of distinct viral clades has been observed in SARS-CoV2 variants across the world and India. Identification of the genomic diversity and the phylodynamic profiles of the prevalent strains of the country are critical to understand the evolution and spread of the variants. We performed whole-genome sequencing of 54 SARS-CoV2 strains collected from COVID-19 patients in Kolkata, West Bengal during August to October 2020. Phylogeographic and phylodynamic analyses were performed using these 54 and other sequences from India and abroad available in GISAID database. Spatio-temporal evolutionary dynamics of the pathogen across various regions and states of India over three different time periods in the year 2020 were analyzed. We estimated the clade dynamics of the Indian strains and compared the clade specific mutations and the co-mutation patterns across states and union territories of India over the time course. We observed that GR, GH and G (GISAID) or 20B and 20A (Nextstrain) clades were the prevalent clades in India during middle and later half of the year 2020. However, frequent mutations and co-mutations observed within the major clades across time periods do not show much overlap, indicating emergence of newer mutations in the viral population prevailing in the country. Further, we explored the possible association of specific mutations and co-mutations with the infection outcomes manifested within the Indian patients.

Insights from a Pan India Sero-Epidemiological survey (Phenome-India Cohort) for SARS-CoV2

To understand the spread of SARS-CoV2, in August and September 2020, the Council of Scientific and Industrial Research (India), conducted a sero-survey across its constituent laboratories and centers across India. Of 10,427 volunteers, 1058 (10.14%) tested positive for SARS CoV2 anti-nucleocapsid (anti-NC) antibodies; 95% with surrogate neutralization activity. Three-fourth recalled no symptoms. Repeat serology tests at 3 (n=346) and 6 (n=35) months confirmed stability of antibody response and neutralization potential. Local sero-positivity was higher in densely populated cities and was inversely correlated with a 30 day change in regional test positivity rates (TPR). Regional seropositivity above 10% was associated with declining TPR. Personal factors associated with higher odds of sero-positivity were high-exposure work (Odds Ratio, 95% CI, p value; 2{middle dot}23, 1{middle dot}92-2{middle dot}59, 6{middle dot}5E-26), use of public transport (1{middle dot}79, 1{middle dot}43-2{middle dot}24, 2{middle dot}8E-06), not smoking (1{middle dot}52, 1{middle dot}16-1{middle dot}99, 0{middle dot}02), non-vegetarian diet (1{middle dot}67, 1{middle dot}41-1{middle dot}99, 3{middle dot}0E-08), and B blood group (1{middle dot}36,1{middle dot}15-1{middle dot}61, 0{middle dot}001). Impact StatementWidespread asymptomatic and undetected SARS-CoV2 infection affected more than a 100 million Indians by September 2020. Declining new cases thereafter may be due to persisting humoral immunity amongst sub-communities with high exposure. FundingCouncil of Scientific and Industrial Research, India (CSIR)

Increased Detection coupled with Social Distancing and Health Capacity Planning Reduce the Burden of COVID-19 Cases and Fatalities: A Proof of Concept Study using a Stochastic Computational Simulation Model

ObjectiveIn absence of any vaccine, the Corona Virus Disease 2019 (COVID-19) pandemic is being contained through a non-pharmaceutical measure termed Social Distancing (SD). However, whether SD alone is enough to flatten the epidemic curve is debatable. Using a Stochastic Computational Simulation Model, we investigated the impact of increasing SD, hospital beds and COVID-19 detection rates in preventing COVID-19 cases and fatalities. Research Design and MethodsThe Stochastic Simulation Model was built using the EpiModel package in R. As a proof of concept study, we ran the simulation on Kasaragod, the most affected district in Kerala. We added 3 compartments to the SEIR model to obtain a SEIQHRF (Susceptible-Exposed-Infectious-Quarantined-Hospitalised-Recovered-Fatal) model. ResultsImplementing SD only delayed the appearance of peak prevalence of COVID-19 cases. Doubling of hospital beds couldnt reduce the fatal cases probably due to its overwhelming number compared to the hospital beds. Increasing detection rates could significantly flatten the curve and reduce the peak prevalence of cases (increasing detection rate by 5 times could reduce case number to half). ConclusionsAn effective strategy to contain the epidemic spread of COVID-19 in India is to increase detection rates in combination with SD measures and increase in hospital beds. HIGHLIGHTSO_LIIncreased Detection of COVID-19 cases must accompany Social Distancing and Health Capacity Planning to reduce the burden of cases and fatalities. C_LIO_LIInterruptive Social Distancing is an effective alternative to continuous Social Distancing. C_LIO_LIGiven the overwhelming burden of COVID-19 fatalities, there is immediate need of co-ordination with the Private Healthcare Sector. C_LIO_LICOVID-19 cases will be peaking after May, 2020 giving us time for Healthcare Capacity Building in the government and private sector both. C_LI

A Machine Learning Model Reveals Older Age and Delayed Hospitalization as Predictors of Mortality in Patients with COVID-19

ObjectiveThe recent pandemic of novel coronavirus disease 2019 (COVID-19) is increasingly causing severe acute respiratory syndrome (SARS) and significant mortality. We aim here to identify the risk factors associated with mortality of coronavirus infected persons using a supervised machine learning approach. Research Design and MethodsClinical data of 1085 cases of COVID-19 from 13th January to 28th February, 2020 was obtained from Kaggle, an online community of Data scientists. 430 cases were selected for the final analysis. Random Forest classification algorithm was implemented on the dataset to identify the important predictors and their effects on mortality. ResultsThe Area under the ROC curve obtained during model validation on the test dataset was 0.97. Age was the most important variable in predicting mortality followed by the time gap between symptom onset and hospitalization. ConclusionsPatients aged beyond 62 years are at higher risk of fatality whereas hospitalization within 2 days of the onset of symptoms could reduce mortality in COVID-19 patients.