Genetic variability of SARS-CoV-2 in biological samples from patients in Moscow

A. S. Speranskaya; Сперанская А. С.; V. V. Kaptelova; Каптелова В. В.; A. E. Samoilov; Самойлов А. Е.; A. Yu. Bukharina; Бухарина А. Ю.; O. Yu. Shipulina; Шипулина О. Ю.; E. V. Korneenko; Корнеенко Е. В.; V. G. Akimkin; Акимкин В. Г.

doi:10.36233/0372-9311-2020-97-6-1

Genetic variability of SARS-CoV-2 in biological samples from patients in Moscow

Authors: Speranskaya A.S.¹, Kaptelova V.V.¹, Samoilov A.E.¹, Bukharina A.Y.¹, Shipulina O.Y.¹, Korneenko E.V.¹, Akimkin V.G.¹
Affiliations:
1. Central Research Institute of Epidemiology
Issue: Vol 97, No 6 (2020)
Pages: 511-517
Section: ORIGINAL RESEARCHES
Submitted: 08.01.2021
Accepted: 08.01.2021
Published: 08.01.2021
URL: https://microbiol.crie.ru/jour/article/view/933
DOI: https://doi.org/10.36233/0372-9311-2020-97-6-1
ID: 933

Cite item

Full Text

Abstract
Full Text
About the authors
References
Supplementary files
Statistics

Abstract

Currently, a lot of attention is given to SARS-CoV-2 subpopulations and their coexistence with different genomic variants within the same patient. In this study, we performed next-generation whole-genome sequencing and assembly of viruses from samples representing swabs or autopsy specimens obtained from patients diagnosed with СOVID-19, which were initially confirmed by the real-time polymerase chain reaction (Ct = 10.4–19.8). Samples were prepared for sequencing by using the SCV-2000bp protocol. The obtained data were checked for presence of more than one SARS-CoV-2 genetic variants in a sample. Variants of nucleotide substitutions, coverage for each variant, and location of the variable position in the reference genome were detected with tools incorporated in the CLC Genomics Workbench program. In our search for variable nucleotide positions, we assumed that the sample had two genetic variants (not more); the threshold value ≥ 90% was set for probability of the identified variant. Variants represented by less than 20% of the reads in the total coverage were not taken into consideration. The obtained results showed that 5 samples had variability, i.e. they had several genetic variants of SARS-CoV-2. In 4 samples, both of the detected genomic variants differed only in one nucleotide position. The fifth sample demonstrated more substantial differences: a total of 3 variable positions and one three-nucleotide deletion. Our study shows that different genetic variants of SARS-CoV-2 can coexist within the same patient.

Keywords

SARS-CoV-2, next generation sequencing, dual infection, quasispecies

Full Text

Introduction

The COVID-19 pandemic caused by the SARSCoV-2 coronavirus began in Wuhan, China, at the end of December in 2019. In Russia, the peak of the first wave was recorded in the middle of May in 2020; the second wave started rising at the end of August. By 11/11/2020, the number of new COVID-19 cases daily registered in Russia had reached approximately 19.9 thousand (more than 1.8 million confirmed cases countrywide.¹In February 2020, the published data obtained by foreign researchers performing next generation sequencing (NGS) of SARS-CoV-2 genomes showed that in the phylogenetic tree, sequences are grouped into two major clades (lineages/types/genotypes) known as L and S. They are differentiated by single-nucleotide polymorphisms in ORF1ab and ORF8 [1]. By 10/11/2020, genomes belonging to 8 major clades: L, O, S, V, G, GR, GH, and GV had been deposited in the GISAID database.²Out of them, 6 clades were identified among Russian isolates at the end of April 2020 [2]; representatives of 7 clades (except for GV) were deposited in GISAID in November 2020 [3].

The SARS-CoV-2 belongs to RNA viruses characterized by high mutation rates, which, in their turn, lead to evolution of quasispecies (sub-populations) within the same host. Currently, the existing quasispecies for SARS-CoV-2 have been identified and recorded [4–8]; there have been studies based on a large amount of the related data. The work [9] includes a bioinformatic analysis of "raw" NGS data from nearly 4 thousand samples obtained at different laboratories and available in the SRA database³. In addition, the same researchers have performed a bioinformatic analysis of NGS data for RNA isolated from swabs of patients from Switzerland and have found that different variants of the SARS-CoV-2 genome coexist within the same patient. U. Fahnøe et al. Explain the phenomenon by natural genetic diversity caused by rapid evolution of viruses (assuming that some of the genomic variants can be artifacts that evolved during preparation of libraries and sequencing). Indeed, NGS is a recognized technique for assessment of genetic variability of viral populations [10, 11] and is used for confirming the existence of viral quasispecies in a patient’s body [12]. However, it is not easy to differentiate truly single-nucleotide variants (SNV) from errors in sequencing and artifacts of sample preparation, especially if it refers to detection of rarely occurring sub-populations.

The phenomenon when the same patient has concurrently two or more variants of the same virus is known as dual infection or coinfection, if it occurs simultaneously with the first infection or sometime later [13]. This phenomenon has been quite extensively studied among viruses of different families and genera [13][14][15][16][17]. The possibility of coinfection for SARS-CoV-2 is still questionable, though the number of arguments supporting this assumption is gradually increasing. Some authors interpret the heterogeneity found in SARSCoV-2 genome sequences as dual infection. For example, this conclusion was offered in the research work [18] describing the results of sequencing of the fragment (795 bp long) of the gene encoding the viral spike protein, when sample were obtained in Iraq, from 19 patients having obvious symptoms of COVID-19. By using Sanger sequencing, double peaks in chromatograms were detected in each of 19 samples. The authors explain the heterogeneity found in the sequencing results as coinfection, while the high percentage (19/19) is explained by the specifics of the national approach to compliance with sanitary regulation (accidental contamination was not considered). In addition to the above preprint, at the end of September 2020, a case of dual infection with SARS-CoV-2 was reported by another group of authors [19] who referred to the data published by S. Ilmjärv et al. [20]. Our preprint article also provides evidence supporting the possibility of dual SARS-CoV-2 infection: We have described the case when genomes belonging to GR and GH clades were detected in a female 90-year-old patient and when the dominant strain changed during the disease course [21].

Mutations occurring in the population and resulting from natural evolution of the virus, similar to infection with another strain, can affect the immune response of the host and change the clinical course of the disease. Some researchers assume that individual mutations of SARS-CoV-2 or their combinations can affect the viral replication rate and speed of disease transmission, cause problems during treatment due to developing antiviral drug resistance mutations [22][23].

The purpose of this study is to demonstrate viral quasispecies at least in part of biological samples from Moscow patients with confirmed SARS-CoV-2 infection.

Materials and methods

The study was conducted on swabs (19 samples) and autopsy material (2 samples) in transport media from patients with ARVI symptoms or with suspected COVID-19, which had been initially delivered to the Department of Molecular Diagnostic Methods at the Central Research Institute of Epidemiology and had been identified as SARS-CoV-2 positive. The identification involved the real-time polymerase chain reaction that was performed with an AmpliSens® Cov-Bat-FL reagent kit (AmpliSens, Russia) in accordance with the user manual. All the samples used in the study contained viral RNA at high concentration (Ct = 10.4–19.8).

The reverse transcription reaction was conducted by using 10 µl of RNA samples and Reverta-L kit (AmpliSens, Russia) in accordance with the user manual. The obtained cDNA was used as a template for amplification of genomic fragments. The SARS-CoV-2 genome fragments were amplified by using the SCV200bp primer panel of our design [25].

The samples were prepared for sequencing in accordance with the protocol [26]. The Q5 High-Fidelity DNA Polymerase (New England BioLabs) was used for amplification of genomic fragments, 35 cycles of amplification. The same polymerase was used for preparation of libraries, 8 cycles of amplification. NGS was completed by using the Illumina HiSeq 1500 platform as well as HiSeq PE Rapid Cluster Kit v2 and HiSeq Rapid SBS Kit v2 reagent kits (500 cycles) or the Illumina MiSeq platform and a MiSeq Reagent Kit v2 reagent kit (500 cycles).

The obtained reads were filtered by quality with the help of Trimmomatic [27]; the sequences of primers were removed by using cutadapt software [28]; the obtained reads were mapped onto the reference sequence EPI_ISL_402124 by using the bowtie2 tool [29]; the SAMtools [30] was used to remove chimeric reads and to receive bam-files. Consensus sequences were obtained with the help of BEDtools [31]. The built-in tools of the CLC Genomics Workbench 8.5 program were used for identification and estimation of the SNV coverage [32]. During the analysis, we proceeded from the assumption that the sample can concurrently have not more than 2 genomic variants (we used the fixed ploidy variant = 2). We set the coverage at ≥ 20% as a threshold value for SNV detection (substitutions with low coverage were disregarded).

Results

The consensus sequences of genomes from the studied samples were deposited in the GISAID database, including SNV in the degenerate nucleotide code format (accession numbers are given in Table 1), except for one sample - d186dl477 (see below for details).

Table 1. Description of degenerate positions in analyzed samples

Sample	Ct	Reads after filtration	Nucleotide position	Total coverage	Nucleotide variants	Variant coverage	Percent of reads	Protein mutation	GISAID accession number
d186s128	16	7 575603	23613	450	C	207	46%	Spike A684V	EPI_ISL_660437
					T	243	54%	Spike A684V
d186I325	13,6	6 337 032	No degenerate positions						EPI_ISL_610232
d186s56	10,4	6 481 901	No degenerate positions						EPI_ISL_610240
d186s51	13,7	12790208	No degenerate positions						EPI_ISL_610239
d186s137	16,3	7 992 766	23916	15388	T	9314	60,53%	Spike V785A	EPI_ISL_660438
					c	6065	39,41%	Spike V785A
d186s144	11.2	6 928 932	No degenerate positions						EPI_ISL_610241
d186s149	13,5	7 026 576	No degenerate positions						EPI_ISL_610242
d186s155	12,6	6 815104	No degenerate positions						EPI_ISL_610243
d186s434	14	7 722 664	No degenerate positions						EPI_ISL_610244
d186dl199	15,1	3 163 502	No degenerate positions						EPI_ISL_610227
d186dl222	18,1	2 885 286	No degenerate positions						EPI_ISL_610228
d186dl240	19,8	1 467 205	8139	24	c	8	33,33%	NSP3S1807F	EPI_ISL_660433
					T	16	66,67%	NSP3S1807F
d186dl282	13,4	4 04 3 559	No degenerate positions						EPI_ISL_610229
d186dl290	13	557 624	24794	4424	G	1596	36,08%	Spike	EPI_ISL_660434
					T	2825	63,86%	A1078S
d186dl294	12,9	6 481 228	No degenerate positions						EPI_ISL_610230
d186dl302	12,4	3 742 501	No degenerate positions						EPI_ISL_610231
d186dl365	14,3	5625499	No degenerate positions						EPI_ISL_610233
d186dl381	13,4	3 066 646	No degenerate positions						EPI_ISL_610234
d186dl389	13	3 424 997	No degenerate positions						EPI_ISL_610235
d186dl441	14,3	4 940 467	No degenerate positions						EPI_ISL_610236
d186dl477	12	2 500 575	21588	17058	T	8950	52,47%	Spike P9L	Nucleotide variants with major coverage — EPI_lSL_660435, with minor coverage — EPI_ISL_660436
					c	8107	47,53%	Spike P9L
			21991- 21993	14810	TTA	11250	75,96%	Spike Y145del
			21991- 21993		—(del)	3544	23,93%	Spike Y145del
			27247	7923	C	4870	61,47%	ORF6 synonymous
					T	3043	38,41%	ORF6 synonymous
			29250	1245	T	873	70,12%	N P326L
					c	370	29,72%	N P326L

After quality filtering and PCR primer trimming, the amount of information per each sample ranged from 0.557 million to 11.965 million reads (the median - 6 million). Then SNVs were analyzed in the viral population for each sample. The mapping of reads onto the reference genome hCoV-19/Wuhan/WIV04/2019 (the GISAID accession number EPI_ISL_402124) showed that some samples had variable/degenerate SNV positions. The proportion of minor subpopulations (estimated by the minor SNV coverage) ranged from 24% to 46%. The presence of variability did not depend on the number of reads or on the viral load in the sample (see samples d186s56, Ct = 10.4 or d186s144, Ct = 11.2, in which no variable positions were found).

SNVs in gene encoding S-protein (Spike Glycoprotein) were detected in 4 samples: d186s128, d186s137, d186dl290, and d186dl477. In the meantime, in one of the samples, the same spike glycoprotein gene had a heterogeneous section represented by the TTA/--- deletion in one of the two genomic variants. In addition, SNVs were found at the site of the ORF1ab gene encoding the nsp3 protein (see sample d186dl240) as well as in genes N and ORF6 (encoding same-name proteins; see sample d186dl477).

Sample d186dl477 is especially noteworthy, as 3 SNVs and a three-nucleotide deletion were detected in it. One of the SNVs leads to synonymous substitution in ORF6; the other cause mutations in the spike protein (P9L and Y145del) as well as in the N protein (P326L). All the detected mutations belong to the category of rare or new mutations (Table 2). The coverage of minor mutation variants in this sample ranged from 29.7% to 47%. We deposited two sequences in the GISAID database: Sequence [EPI_ISL_660435] includes major mutation variants, while sequence [EPI_ISL_660436] includes minor variants.

Table 2. Characterization of mutations found in sample d186dl477

Protein	EPI_ ISL_660435	EPI_ISL_660436	Frequency of found mutations among known SARS-CoV-2 sequences according to GISAID data, as of November 27, 2020
Protein	EPI_ ISL_660435	EPI_ISL_660436	%	number of genoms	number of countries
NSP2	T547I		0,07	55	10
NSP3	V1847F		>0,01	7	2
NSP12	P323L		88,88	184 070	125
Spike	P9L	-	0,01	27	7
	-	Y145del	Found for the first time
	D614G		89,28	18 5187	26
N	R203K		36,46	75 504	104
	G204R		36,15	74 866	103
	M210I		0,24	488	18
	A211V		0,08	175	22
	P326L	—	0,02	41	8

Note. Mutations for which the frequency of occurrence in the world is ≤ 0.25 are in bold.

Discussion

In our study, we intentionally narrowed down our search to strongly represented variants, as our main objective was to show wide occurrence of heterogeneous populations of SARS-CoV-2. We analyzed the NGS data for SARS-CoV-2 genomes and evaluated the representation of viral subpopulations by using variability search algorithms implemented as built-in tools of the CLC Genomics Workbench program. We did not try to find optimum criteria for SNV identification, being satisfied by using rigid criteria (min 20% of the total coverage of the analyzed position at detection reliability of at least 90%). We identified SNVs in 5 samples out of 21. In 4 samples, quasispecies differed in occasional SNVs.

RNA viruses are characterized by high mutation rates, which frequently cause developing of quasispecies within the same host. Numerous studies show that several viruses with different SNV in genomes have been found concurrently existing in samples from COVID-19 patients [8][9][10][11][21]. We assume that occasional SNVs found in 4 samples can be also explained by natural evolution of viral genomes in the host.

The fifth sample (d186dl477) differed from other heterogeneous samples by 3 SNPs and 1 heterogeneous three-nucleotide section that were detected in it. However, the values of relative coverage in these positions did not show any substantial difference, thus leading to the conclusion that the sample has a combination of strains. We assume that this phenomenon can be explained in two ways: unprecedentedly fast evolution of the virus in the body of this patient or infection with different strains of SARS-CoV-2.

Recent publications show that the mutation rate for SARS-CoV-2 is almost identical to the mutation rate of the SARS-CoV genome (0.80–2.38 × 10^–3 nucleotide substitutions per site per year) [19][30][31]. Having statistically analyzed a large amount of "raw" sequencing data from different laboratories, J. Kuipers et al. [11] demonstrated that heterogeneity of the viral population in a sample may correlate with the age of the patient. Sample d186dl477 was obtained from an 84-year-old female patient. If in our theoretical calculations, we assume that the mutation rate equals the highest possible rate (2.38 × 10^–3 nucleotide substitutions per site per year), then up to 10 mutations can evolve in the SARS-CoV-2 genome during 5 days of disease progression.

The idea of evolution is also supported by the fact that four mutations out of the mutations common to both strains are characterized by low occurrence worldwide — >0.01–0.24%. The SNV occurrence resulting from the consequential evolution should have affected the mutation rate. However, the observed difference in the coverage values for minor SNVs is not significant. In the absence of clinical data and information about the duration of the disease in the patient we obtained sample d186dl477 from, we cannot decidedly assert that heterogeneity is a consequence of the natural evolution of the virus. A lack of data gives no support for the alternative assumption (i.e. coinfection resulting from the infection with the second strain of SARS-CoV-2).

Tang X., Wu C., Li X., Song Y., Yao X., Wu X., et al. On the origin and continuing evolution of SARS-CoV-2. Natl. Sci. Rev. 2020; 7(6): 1012–23. https://doi.org/10.1093/nsr/nwaa036
Komissarov A.B., Safina K.R., Garushyants S.K., Fadeev A.V., Sergeeva M.V., Ivanova A.A., et al. Genomic epidemiology of the early stages of SARS-CoV-2 outbreak in Russia. medRxiv. Preprint. 2020. https://doi.org/10.1101/2020.07.14.20150979
Sýkorová E., Fajkus J., Mezníková M., Lim K.Y., Neplechová K., Blattner F.R., et al. Minisatellite telomeres occur in the family Alliaceae but are lost in Allium. Am. J. Bot. 2006; 93(6): 814–23. https://doi.org/10.3732/ajb.93.6.814
Nyayanit D., Yadav P.D., Kharde R., Shete-Aich A. Quasispecies analysis of the SARS-CoV-2 from representative clinical samples: A preliminary analysis. Indian J. Med. Res. 2020; 152(1): 105. https://doi.org/10.4103/ijmr.ijmr_2251_20
Jary A., Leducq V., Malet I., Marot S., Klement-Frutos E., Teyssou E., et al. Evolution of viral quasispecies during SARSCoV-2 infection. Clin. Microbiol. Infect. 2020; 26(11): 1560. e1-1560.e4. https://doi.org/10.1016/j.cmi.2020.07.032
Chaudhry M.Z., Eschke K., Grashoff M., Abassi L., Kim Y., Brunotte L., et al. SARS-CoV-2 quasispecies mediate rapid virus evolution and adaptation. bioRxiv. Preprint. 2020. https://doi.org/10.1101/2020.08.10.241414
Xu D., Zhang Z., Wang F.S. SARS-associated coronavirus quasispecies in individual patients. N. Engl. J. Med. 2004; 350(13): 1366–7. https://doi.org/10.1056/nejmc032421
Park D., Huh H.J., Kim Y.J., Son D.S., Jeon H.J., Im E.H., et al. Analysis of intrapatient heterogeneity uncovers the microevolution of Middle East respiratory syndrome coronavirus. Mol. Case Stud. 2016; 2(6): a001214. https://doi.org/10.1101/mcs.a001214
Kuipers J., Batavia A.A., Jablonski K.P., Bayer F., Borgsmüller N., Dondi A., et al. Within-patient genetic diversity of SARS-CoV-2. bioRxiv. Preprint. 2020. https://doi.org/10.1101/2020.10.12.335919
McElroy K., Zagordi O., Bull R., Luciani F., Beerenwinkel N. Accurate single nucleotide variant detection in viral populations by combining probabilistic clustering with a statistical test of strand bias. BMC Genomics. 2013; 14(1): 501. https://doi.org/10.1186/1471-2164-14-501
Fahnøe U., Pedersen A.G., Dräger C., Orton R.J., Blome S., Höper D., et al. Creation of functional viruses from non-functional cDNA clones obtained from an RNA virus population by the use of ancestral reconstruction. PLoS One. 2015; 10(10): e0140912. https://doi.org/10.1371/journal.pone.0140912
Kireev D.E., Lopatukhin A.E., Murzakova A.V., Pimkina E.V., Speranskaya A.S., Neverov A.D., et al. Evaluating the accuracy and sensitivity of detecting minority HIV-1 populations by Illumina next-generation sequencing. J. Virol. Methods. 2018; 261: 40–5. https://doi.org/10.1016/j.jviromet.2018.08.001
Лаповок И.А., Лопатухин А.Э., Киреев Д.Е. Двойная ВИЧ-инфекция: эпидемиология, клиническая значимость и диагностика. Инфекционные болезни. 2019; 17(2): 81–7. [Lapovok I.A., Lopatukhin A.E., Kireev D.E. Dual HIV infection: epidemiology, clinical significance, and diagnosis. Infektsionnye bolezni. 2019; 17(2): 81–7. (in Russ.)] https://doi.org/10.20953/1729-9225-2019-2-81-87
Falchi A., Arena C., Andreoletti L., Jacques J., Leveque N., Blanchon T., et al. Dual infections by influenza A/H3N2 and B viruses and by influenza A/H3N2 and A/H1N1 viruses during winter 2007, Corsica Island, France. J. Clin. Virol. 2008; 41(2): 148–51. https://doi.org/10.1016/j.jcv.2007.11.003
Semple M.G., Cowell A., Dove W., Greensill J., McNamara P.S., Halfhide C., et al. Dual infection of infants by human metapneumovirus and human respiratory syncytial virus is strongly associated with severe bronchiolitis. J. Infect. Dis. 2005; 191(3): 382–6. https://doi.org/10.1086/426457
van der Kuyl A.C., Cornelissen M. Identifying HIV-1 dual infections. Retrovirology. 2007; 4: 67. https://doi.org/10.1186/1742-4690-4-67
Weinberg A., Bloch K.C., Li S., Tang Y.W., Palmer M., Tyler K.L. Dual infections of the central nervous system with Epstein‐Barr virus. J. Infect. Dis. 2005; 191(2): 234–7. https://doi.org/10.1086/426402
Hashim H.O., Mohammed M.K., Mousa M.J., Abdulameer H.H., Alhassnawi A.T., Hassan S.A., et al. Unexpected co-infection with different strains of SARS-CoV-2 in patients with COVID-19. Preprints.org. 2020. Preprint. https://doi.org/10.20944/preprints202009.0375.v1
Liu S., Shen J., Fang S., Li K., Liu J., Yang L., et al. Genetic spectrum and distinct evolution patterns of SARS-CoV-2. Front. Microbiol. 2020; 11: 593548. https://doi.org/10.3389/fmicb.2020.593548
Gudbjartsson D.F., Helgason A., Jonsson H., Magnusson O.T., Melsted P., Norddahl G.L., et al. Spread of SARS-CoV-2 in the Icelandic population. N. Engl. J. Med. 2020; 382(24): 2302–15. https://doi.org/10.1056/nejmoa2006100
Samoilov A., Kaptelova V.V., Bukharina A.Y., Shipulina O.Y., Korneenko E.V., Lukyanov A.V. et al. Change of dominant strain during dual SARS-CoV-2 infection: preprint. medRxiv. Preprint. 2020. https://doi.org/10.1101/2020.11.29.20238402
Ilmjärv S., Abdul F., Acosta-Gutiérrez S., Estarellas C., Galdadas I., Casimir M., et al. Epidemiologically most successful SARS-CoV-2 variant: concurrent mutations in RNA-dependent RNA polymerase and spike protein. medRxiv. Preprint. 2020. https://doi.org/10.1101/2020.08.23.20180281
Speranskaya A., Kaptelova V., Valdokhina A., Bulanenko V., Samoilov A., Korneenko E., et al. SCV-2000bp: a primer panel for SARS-CoV-2 full-genome sequencing. bioRxiv. Preprint. 2020. https://doi.org/10.1101/2020.08.04.234880
Kaptelova V.V., Speranskaya A.S. Protocol for SCV-2000bp: a primer panel for SARS-CoV-2 full-genome sequencing v1. https://dx.doi.org/10.17504/protocols.io.bn77mhrn
Bolger A.M., Lohse M., Usadel B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics. 2014; 30(15): 2114–20. https://doi.org/10.1093/bioinformatics/btu170
Martin M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet.journal. 2011; 17(1): 10. https://doi.org/10.14806/ej.17.1.200
Langmead B., Salzberg S.L. Fast gapped-read alignment with Bowtie 2. Nat. Methods. 2012; 9(4): 357–9. https://doi.org/10.1038/nmeth.1923
Li H., Handsaker B., Wysoker A., Fennell T., Ruan J., Homer N., et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics. 2009; 25(16): 2078–9. https://doi.org/10.1093/bioinformatics/btp352
Quinlan A.R., Hall I.M. BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics. 2010; 26(6): 841–2. https://doi.org/10.1093/bioinformatics/btq033
Romano M., Ruggiero A., Squeglia F., Maga G., Berisio R. A structural view of SARS-CoV-2 RNA replication machinery: RNA synthesis, proofreading and final capping. Cells. 2020; 9(5): 1267. https://doi.org/10.3390/cells9051267
Zhao Z., Li H., Wu X., Zhong Y., Zhang K., Zhang Y.P., et al. Moderate mutation rate in the SARS coronavirus genome and its implications. BMC Evol. Biol. 2004; 4: 21. https://doi.org/10.1186/1471-2148-4-21