ABSTRACT
Many SARS-CoV-2 variants have emerged during the course of the COVID-19 pandemic. These variants have acquired mutations conferring phenotypes such as increased transmissibility or virulence, or causing diagnostic, therapeutic, or immune escape. Detection of Alpha and the majority of Omicron sublineages by PCR relied on the so-called S gene target failure due to the deletion of six nucleotides coding for amino acids 69-70 in the spike (S) protein. Detection of hallmark mutations in other variants present in samples relied on whole genome sequencing. However, whole genome sequencing as a diagnostic tool is still in its infancy due to geographic inequities in sequencing capabilities, higher cost compared to other molecular assays, longer turnaround time from sample to result, and technical challenges associated with producing complete genome sequences from samples that have low viral load and/or high background. Hence, there is a need for rapid genotyping assays. In order to rapidly generate information on the presence of a variant in a given sample, we have created a panel of four triplex RT-qPCR assays targeting 12 mutations to detect and differentiate all five variants of concern: Alpha, Beta, Gamma, Delta and Omicron. We also developed an expanded pentaplex assay that can reliably distinguish among the major sublineages (BA.1-BA.5) of Omicron. In silico, analytical and clinical testing of the variant panel indicate that the assays overall exhibit high sensitivity and specificity. This variant panel can be used as a Research Use Only screening tool for triaging SARS-CoV-2 positive samples prior to whole genome sequencing.
ABSTRACT
We previously interrogated the relationship between SARS-CoV-2 genetic mutations and associated patient outcomes using publicly available data downloaded from GISAID in October 2020 [1]. Using high-level patient data included in some GISAID submissions, we were able to aggregate patient status values and differentiate between severe and mild COVID-19 outcomes. In our previous publication, we utilized a logistic regression model with an L1 penalty (Lasso regularization) and found several statistically significant associations between genetic mutations and COVID-19 severity. In this work, we explore the applicability of our October 2020 findings to a more current phase of the COVID-19 pandemic. Here we first test our previous models on newer GISAID data downloaded in October 2021 to evaluate the classification ability of each model on expanded datasets. The October 2021 dataset (n=53,787 samples) is approximately 15 times larger than our October 2020 dataset (n=3,637 samples). We show limitations in using a supervised learning approach and a need for expansion of the feature sets based on progression of the COVID-19 pandemic, such as vaccination status. We then re-train on the newer GISAID data and compare the performance of our two logistic regression models. Based on accuracy and Area Under the Curve (AUC) metrics, we find that the AUC of the re-trained October 2021 model is modestly decreased as compared to the October 2020 model. These results are consistent with the increased emergence of multiple mutations, each with a potentially smaller impact on COVID-19 patient outcomes. Bioinformatics scripts used in this study are available at https://github.com/JPEO-CBRND/opendata-variant-analysis. As described in Voss et al. 2021, machine learning scripts are available at https://github.com/Digital-Biobank/covid_variant_severity.
ABSTRACT
BackgroundEmerging and reemerging infectious diseases such as the novel Coronavirus disease, COVID-19 and Ebola pose a significant threat to global society and test the public health communitys preparedness to rapidly respond to an outbreak with effective diagnostics and therapeutics. Recent advances in next generation sequencing technologies enable rapid generation of pathogen genome sequence data, within 24 hours of obtaining a sample in some instances. With these data, one can quickly evaluate the effectiveness of existing diagnostics and therapeutics using in silico approaches. The propensity of some viruses to rapidly accumulate mutations can lead to the failure of molecular detection assays creating the need for redesigned or newly designed assays. ResultsHere we describe a bioinformatics system named BioLaboro to identify signature regions in a given pathogen genome, design PCR assays targeting those regions, and then test the PCR assays in silico to determine their sensitivity and specificity. We demonstrate BioLaboro with two use cases: Bombali Ebolavirus (BOMV) and the novel Coronavirus 2019 (SARS-CoV-2). For the BOMV, we analyzed 30 currently available real-time reverse transcription-PCR assays against the three available complete genome sequences of BOMV. Only two met our in silico criteria for successful detection and neither had perfect matches to the primer/probe sequences. We designed five new primer sets against BOMV signatures and all had true positive hits to the three BOMV genomes and no false positive hits to any other sequence. Four assays are closely clustered in the nucleoprotein gene and one is located in the glycoprotein gene. Similarly, for the SARS-CoV-2, we designed five highly specific primer sets that hit all 145 whole genomes (available as of February 28, 2020) and none of the near neighbors. ConclusionsHere we applied BioLaboro in two real-world use cases to demonstrate its capability; 1) to identify signature regions, 2) to assess the efficacy of existing PCR assays to detect pathogens as they evolve over time, and 3) to design new assays with perfect in silico detection accuracy, all within hours, for further development and deployment. BioLaboro is designed with a user-friendly graphical user interface for biologists with limited bioinformatics experience.