This document presents a next-generation sequencing (NGS)-based method for diagnosing COVID-19 using a single-step RNA extraction. The method was tested on 336 clinical samples and showed high accuracy compared to RT-PCR, the gold standard method. Key advantages of the NGS approach include higher scalability allowing thousands of samples to be tested simultaneously, and higher sensitivity to detect SARS-CoV-2 RNA compared to RT-PCR. The method provides individual qualitative results for each patient sample along with viral load estimates based on sequencing coverage.

1. Presented by:
Mesele Tilahun Belete
(Integ.2nd year)
Advisor: MoonJae Sun
Apr 22, 2021
KRISS, UST, S.Korea

2. Contents
1. INTRODUCTION
2. METHOD AND MATERIALS
3. RESULT AND DISCUSSION
5. Conclusion

3. Introduction to the Virus
 COVID-19 Pandemic: One of the most serious new health threats in the modern history
of humanity.
In March 2020, the WHO declared COVID-19 as pandemic,
- issued guidelines about clinical and epidemiological findings
 COVID-19 is caused by SARS-CoV-2, a Beta-coronavirus closely related to the SARS
virus.
There after, 2019-nCOV was named SARS-CoV-2.
Transmission: Respiratory, Naso-pharyngeal and Speech droplets by direct inoculation
via touching of fomites or breathing in such droplets. Asymptomatic carriage.

4. COVID19
 family / genus Coronaviridae , Betacoronavirus,
Enveloped, spherical, Monopartite, linear ssRNA(+) genome of 26-32kb in size
(the largest of all RNA virus genomes).
 Capped, and polyadenylated.
genome organization and expression, usually composed by
- 10 or more nonstructural proteins (nsp proteins) and
-structural proteins spike (S)
-envelope (E)
-membrane (M) and
-nucleocapsid(N)

5. Genome organization
https://viralzone.expasy.org/764
GENE EXPRESSION
The virion RNA is infectious and serves as both the genome and viral messenger
RNA. gRNA encodes ORF1a, as for ORF1b, it is translated by ribosomal
frameshifting. Resulting polyproteins pp1a and pp1ab are processed into the viral
polymerase (RdRp) and other ns proteins involved in RNA synthesis.
- Structural proteins are expressed as subgenomic RNAs.

6. Common Diagnostic Methods for RNA viruses
The diagnostic assays for
RNA viruses are classified
into 5 major categories:
https://aip.scitation.org/doi/10.1063/5.0021554

7. Statement of problem
 RT-PCR assays- gold-standard for molecular diagnosis of COVID19. Because of the
rising number of cases and rapid spreading
 Shortage of RT-PCR supplies, especially the ones involved in RNA extraction
 Highly dependent on equipment that usually testing of 96 samples at a time
 Author proposed a cost-effective molecular NGS-based test for diagnosis of COVID19,
which uses a single-step RNA extraction and presents high scalability and accuracy
when compared to the gold-standard RT-qPCR.

8. COVID19- NGS technology
 The massively parallel sequencing technology known as next-generation
sequencing (NGS) has revolutionized the biological sciences.
 With its ultra-high throughput, scalability, and speed,
sequence hundreds and thousands of genes or whole genome in a short
period of time.
 A single run of the NGS-based test using the Illumina NextSeq 550 mid-end
sequencing equipment is able to multiplex 1,536 patient’s samples, providing
individual semi-qualitative results (detected, not detected).
 Usage of the high-end Illumina Novaseq platform may yield diagnostic for
up to 6144 samples in a single run.

9. Methods
Sample Collection
Analyzing
Library generation
RT-PCR
RNA extraction & Internal control
Sequencing
purification & normalization

10. Method cont’d
MS2 phage RNA spike- added as internal control of true –ve samples
i.e., -ve do not yield MS2 seq after bioinformatics analyses
QuickExtract DNA Extraction Solution- This requires only heat
treatment for 8 min direct to a plaque where samples are placed
MagMAX Viral (Pathogen II NA) Isolation Kit- semi-automated
procedure taking about 90 min to extract 96 samples
A total of 336 samples with previous results for COVID19 RT-
qPCR were de-identified and selected from routine testing.
The samples used processed in a period shorter than 24 h
Clinical samples were collected directly from patients for diagnosis.
Sample
Collection
RNA extraction
Internal control

11. Amplicon
generation
performed individually using SuperScript™ III One-Step RT-PCR
combination of multiplex primers designed to amplify highly
conserved regions of SARS-CoV-2 and MS2 control genomes.
Library
generation
Each patient’s set of individually amplified fragments for SARS-CoV-2
and MS2 genomes was combined in a pool using the Biomek FX
After pool combination, libraries were built using KAPA HiFi HotStart
Enzyme (Roche) and Illumina Nextera XT v2 Index kit.
Method cont’d

12. Library
purification
normalization
Library purification- performed in the Biomek FX Automation
(Beckman Coulter) using Agencourt AM Pure XP
For normalization, all libraries were quantified by Qubit Fluorometer
using Qubit™ dsDNA HS Assay Kit and
size assessment was performed on Agilent 2200 TapeStation system .
Method cont’d

13. General scheme of how samples are processed in PCR and library plates
(A) Each sample is individually
amplified with a multiplex
combination of primers for
two SARS-CoV-2 targets and
one MS2 control.
(B) Every four columns of the PCR
plate will be transformed into one
column for the library plate.
(C) Four complete PCR plaques (384
samples) generate one library plaque (96
individual sequencing libraries).

14. Sequencing- Sample pools were diluted to 2 nM based on the Qiaseq
Library Quant Assay Kit (Qiagen) measurements and the size
information was performed on Agilent 2200 TapeStation system.
Sequencing
Analyses
developed an online platform termed VarsVID, as it is a feature of the
Varstation platform for genetic analyses (https://varstation.com/).
General steps of the bioinformatics pipeline that is executed in the EC2 instance is:
(1) Quality Control,
(2) Mapping to references
(3) Assessing reads aligned to targets and
(4) Demultiplexing to generate patient’s individual qualitative results.
to evaluate the r/p b/n CT and FPM values, a piecewise linear regression model was fit
FPM values were log-transformed, since it is in principle a linear measure of viral load,
while CT is a logarithmic measure.
The analysis was carried out on PyMC3, an open-source Bayesian inference framework
Method cont’d

15. Result and Discussion
-developed a complete workflow to execute the NGS-based testing, including wet-lab and bioinformatics analyses, as shown in Fig. 1

16. Performance and limit of detection
 In this work, authors showed a new amplicon-based methodology to amplify
highly conserved regions of the SARS-CoV-2 genome and to uniquely identify
these fragments using NGS sequencing.
 Two types of RNA extraction used to test NGS protocol:
1. Simple one-step RNA extraction with QuickExtract (Lucigen, WI, USA) and
2. Regular kit extraction using MagMAX Viral/Pathogen II Nucleic Acid Isolation
Kit (MVP II; Thermo Fisher).

17. Performance and limit of detection
A total of 269 clinical samples extracted with Quick Extract and tested using the NGS approach
RT-qPCR
/NGS
Quick extract extraction Thermo fisher extraction
Positive Negative Total Positive Negative Total
Positive 102 10 112 198 16 224
Negative 1 156 157 2 62 64
Total 103 166 269 200 78 278
Rows are actual RT-PCR
results and columns are
NGS results.
Table 1 Contingency table showing results in absolute count for each extraction method.
Performance
test
NGS
Quick Extract for
all samples
Quick Extract for
samples with CT <30
Thermo Fisher
Accuracy 99.5 98.1 93.5
Sensitivity 99.1 96.1 96.9
Specificity 99.4 99.4 92.5
PPV 99 99 79.5
NPV 94 97.1 99
Based on these results author also calculated the positive predictive value (PPV) and the negative predictive value (NPV).
PPV value was 99% (102 NGS true positive and 1 NGS false positive) and NPV value was 94% (156 true negatives and 10 false negatives).
author noted that most false negative results presented RT-qPCR Cycle Threshold (CT) values higher than 30 (6 false negatives).
269 sample 
1. 112sample +ve RT-qPCR
102 NGS confirmed
2. 156 sample--ve RT-qPCR
155 NGS confirmed

18. Performance results for the NGS approach in three different scenarios:

19. NGS Coverage
 NGS coverage was assessed for all positive samples
 A measure of fragments per million (FPM)- suitable to compare coverage
among different samples and sequencing runs.
The r/p b/n CT & FPM, evaluated by piecewise linear regression model fit.
FPM values were log-transformed, since it is in principle a linear measure
of viral load, while CT is a logarithmic measure.
The analysis was carried out on PyMC3, an open-source Bayesian inference
framework.

20. NGS coverage
Correlation between RT-PCR cycle threshold and Fragments per Million.
The blue line indicates the posterior median of the Ct for each value of FPM.
Blue shade indicates 95% posterior predictive interval (that is, 95% of new samples are expected to fall within this range)

21.  16 non-agreements observed as false positives when a regular kit approach used.
• However, NGS-based technique combined with regular kit extraction being more
sensitive to detect SARS-CoV-2 RNA than RT-qPCR.
• In SARS‐CoV‐2 infections, the combination of several methods improves the
diagnostic efficiency
• Author tested 2 of the 16 alleged false positives using a more sensitive RT-qPCR
technique (Xpert Xpress SARS-CoV-2, Cepheid) and both results came positive
with CT values above 40.
• Therefore, authors conjecture that the NGS technique presented in this work has a
limit of detection of fewer copies/reaction than gold standard RT-qPCR.

22. Conclusion
 NGS-based technique combined with regular kit extraction being more sensitive to
detect SARS-CoV-2 RNA than RT-qPCR.
• In SARS‐CoV‐2 infections, the combination of several methods improves the
diagnostic efficiency
It is important to aligned with other alternative solutions for COVID19 diagnosis,
such as RT-lamp, mass spectrometry and CRISPR based diagnostic.
• Detected results are provided with fragments per million (FPM) values, which was
demonstrated to correlate with RT-qPCR Cycle Threshold (CT) values.
• Performance results when compared with RT-qPCR show general accuracy of 96%,
and 98% when only samples with CT values (gene N) lower than 30 are considered.

23. Conclusion
 Author developed an online platform, termed VarsVID, to help test executors to easily
scale testing numbers (https://varstation.com/).
• Sample registering, wet-lab worksheets generation, sample sheet for sequencing and
results’ display are all features provided by VarsVID.
• Altogether, these results will contribute to control COVID19 pandemics.