Recombination-aware phylogenetic analysis sheds light on the evolutionary origin of SARS-CoV-2

doi:10.21203/rs.3.rs-2189993/v1

Download PDF

Article

Recombination-aware phylogenetic analysis sheds light on the evolutionary origin of SARS-CoV-2

https://doi.org/10.21203/rs.3.rs-2189993/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 04 Jan, 2024

Read the published version in Scientific Reports →

You are reading this latest preprint version

SARS-CoV-2 can infect human cells through the recognition of the human angiotensin-converting enzyme 2 (ACE2) receptors. This affinity is given by six amino acid located in the receptor binding domain (RBD) region within the Spike protein. Genetic recombination involving bat and pangolin Sarbecoviruses, and natural selection have been proposed as possible explanations for the acquisition of these amino acids. In this study we employed Bayesian phylogenetics to jointly reconstruct the phylogeny of the RBD among human, bat and pangolin Sarbecoviruses and detect recombination events affecting this region of the genome. A recombination event involving RaTG13, the closest relative of SARS-CoV-2 that lacks five of the six residues, and an unsampled Sarbecovirus lineage was detected. This result suggests that the key amino acids were likely present in the common ancestor of SARS-CoV-2 and RaTG13, with the latter losing five of the amino acids as the result of recombination.

Biological sciences/Evolution/Phylogenetics

Biological sciences/Microbiology/Virology/Sars cov 2

At the end of the year 2019, human cases of a new respiratory disease were detected in Wuhan, China^1,2. It was determined that the infectious agent was a new RNA virus member of the Family Coronaviridae, named Severe acute respiratory syndrome coronavirus 2 or SARS-CoV-2³. The virus spread worldwide during the next months, and cases have been reported in more than 200 countries. Over 600 million infections and more than 6.5 million deaths have been attributed to this virus as of September 2022 (https://covid19.who.int/). SARS-CoV-2 is the third coronavirus capable of inducing severe respiratory diseases in humans to emerge in the last 18 years, after the severe acute respiratory syndrome coronavirus (SARS-CoV), detected in China in 2002⁴, and the Middle East respiratory syndrome coronavirus (MERS-CoV) identified in the year 2012 in Saudi Arabia⁵.

SARS-CoV-2, along with SARS-CoV, is a member of the subgenus Sarbecovirus, and has a positive sense single-stranded RNA genome with a length of approximately 30,000 base pairs that encodes four major structural proteins: spike, envelope, membrane, and nucleocapsid, encoded by the S, E, M and N genes respectively⁶. The spike protein has a central role in cell infection and pathogenesis, since it mediates the recognition of cellular receptors and the binding of the viral and cell membranes, a process that ultimately leads to the entry of the virus into the cell⁷. The spike protein contains a Receptor Binding Domain (RBD), which gives the virus an affinity for the angiotensin-converting enzyme 2 (ACE2), a human cell receptor also used by SARS-CoV for the binding process^1,8. The RBD is the most variable part of the genome among coronaviruses and has an important role in determining the host range of each viral species^9,10. The RBD of SARS-CoV-2 is characterized by six contact amino acid residues that are essential for the binding to the ACE2 receptors⁹, and are present in a region known as the variable loop¹¹.

The evolutionary origin of SARS-CoV-2 remains unclear so far. Analyses of full genome sequences indicate that the closest known relatives of this virus are the bat Sarbecoviruses RmYN02 and RaTG13^2,12, which suggests an emergence in humans following a spillover from bats directly or through an intermediate host. However, Sarbecoviruses, like other coronaviruses, are highly recombinant^11,13 and the examination of the SARS-CoV-2 RBD sequence has suggested more complex scenarios for its origin involving genetic recombination. Indeed, while bat viruses are the closest relatives of SARS-CoV-2 considering the full genome sequences, the RBDs of SARS-CoV-2 and GD410721, a Sarbecovirus isolated from Malayan pangolins (Manis javanica), are highly similar, and share the six amino acid residues that confer the affinity for ACE2 receptors^14,15.

Hence, three of the four main hypotheses for the origin of the variable loop of SARS-CoV-2 include recombination (Fig. 1). First, it has been postulated that SARS-CoV-2 acquired the variable loop of the RBD after a recombination event with a pangolin Sarbecovirus (Fig. 1a). Previous studies have employed tools for the detection of recombination breakpoints and the generation of similarity plots, to find signals of recombination events in the S gene involving SARS-CoV-2, RaTG13 and GD410721^15–17. A second hypothesis proposed by Boni et al.¹¹ suggests that the common ancestor of SARS-CoV-2, GD410721 and RaTG13 was capable of recognizing the human ACE2 receptors and part of the RBD of RaTG13 was replaced through recombination with another Sarbecovirus missing the ACE2-specific residues (Fig. 1b). The recent discovery of Sarbecoviruses in different species of cave bats (BANAL-52, -103, and − 236), presenting an almost identical RBD to the one from SARS-CoV-2 has motivated a third hypothesis, in which SARS-CoV-2 could have acquired the contact residues as the result of recombination events among bat Sarbecoviruses¹⁸ (Fig. 1c). Finally, the affinity for human ACE2 receptors found in viruses that infect different hosts carrying the same type of receptors could also be the result of convergent evolution⁹ (Fig. 1d).

Given the potential importance of recombination in the evolutionary history of the S gene, the detection of recombination events has become a very important task in evolutionary studies of SARS-CoV-2. The presence of recombination among Sarbecoviruses has been explored previously with tools like SimPlot¹⁹, RDP4²⁰, RDP5²¹ and GARD²². These methods allow the identification of recombinant strains and recombination breakpoints, as well as potential parental strains involved. However, they are unable to estimate Ancestral Recombination Graphs to represent the reticulated evolution induced by genetic recombination. This seriously limits the possibility to generate a comprehensive evolutionary picture, does not allow to propagate phylogenetic uncertainties in the estimation of recombination events^23–25.

The reconstruction of ARGs from sequence data is a notoriously difficult task. Nevertheless, several software packages have been developed to tackle this challenge. One example is the Bayesian approach implemented in the BEAST2 package Bacter²⁴. Bacter is an implementation of the ClonalOrigin model²⁶ which can be used to estimate a special type of ARG referred to as Ancestral Conversion Graphs (ACG)s. ACGs consist of a backbone bifurcating phylogeny representing the evolution of the major part of the genetic material (referred to as the "clonal frame"), together with recombination events involving donor and recipient lineages on the clonal frame (Supplementary Fig. S1 online). The method allows to estimate recombination events within a dated phylogeny together with measures of statistical support and uncertainties. This allows us to date individual recombination events, which in turn provide us with an estimation of the emergence time of recombinant lineages. A different Bayesian approach to perform recombination-aware phylogenetic analyses produces recombination networks, where recombinant edges are also integrated within the phylogeny²⁷. Although unlike Bacter, this approach doesn’t estimate the posterior support for the arrival point of the recombination on the recipient lineage. In this study we employed Bacter to detect recombination events within the RBD region of 39 Sarbecovirus genomes, with the goal of clarifying the origin of the amino acids located in the variable loop that give SARS-CoV-2 the high affinity to the ACE2 receptors of human cells.

Temporal signal and model selection

The initial data set consisted of a sequence alignment of the RBD region of 87 Sarbecoviruses, with potentially misaligned regions masked by the program Gblocks²⁸. Using a Bayesian Evaluation of Temporal Signal (BETS) analysis²⁹ we identified a significant level of temporal signal in our dataset (logBF = 22.6). The BEAST2 package bModelTest³⁰ selected as the best substitution model a modification of the Hasegawa-Kishino-Yano (HKY) model³¹ with three different substitution rates (r_AC=r_AT=r_CG, r_AG=r_CT, and r_GT) and a proportion of invariant sites of 0.3. A model comparison with a Stepping-Stone sampling³² determined the Bayesian Skyline coalescent as the best fitting tree prior (logBF = 30.57).

Bacter analysis and dating of the clonal frame

As a Bacter analysis is computability demanding, a representative subsample of 39 viral sequences was selected with the program Uclust³³. The combined results of three Bacter analyses estimated a median evolutionary rate of 2.96E^− 4 substitutions per site per year (95% high posterior density interval HPDI: 2.05E^− 4 – 5E^− 4), which is comparable to previous estimates¹¹. The maximum clade credibility tree of the clonal frame shows a virus from cave bats, BANAL-103, as the closest relative of SARS-CoV-2, with RaTG13 and GD410721 representing the second and third closest lineages to the human virus respectively (Fig. 2). The root of the tree has an estimated median age of 1661 years before the present (yBP) (95% HPDI: 2019 − 998 yBP). The split between SARS-CoV-2 and BANAL-103 occurred around 131 yBP (95% HPDI: 198 − 65 yBP), while the divergence between these viruses and RaTG13 happened 189 yBP (95% HPDI: 283 − 97 yBP). The common ancestor of the viruses capable of recognizing ACE2 receptors (SARS-CoV-2, BANAL-103 and GD410721), and RaTG13 has a median age of 337 yBP (95% HPDI: 476 − 172 yBP). The current implementation of Bacter supports only the strict molecular clock model, while a previous dating analysis employed a relaxed clock model¹¹. With a relaxed clock model, we may expect changes in the inferred timing of branching and recombination events. The topological placement of recombination events in the phylogeny, however, would not be affected.

Bacter analysis and recombination events

A summary ACG obtained with a posterior probability threshold of 0.5 didn’t detect significant recombination events involving SARS-CoV-2, RaTG13, BANAL-103 or GD410721 (Supplementary Fig. S2 online). While this result supports the host adaptation hypothesis, a summary ACG with a reduced posterior threshold of 0.1 detected four recombination events with overlapping recombinant regions and with arrival points, which indicate the recipient of the recombinant region, located in the terminal branch leading to RaTG13 (Supplementary Fig. S3 online). The four donor sequences are also unsampled members of the same phylogenetic clade, and one is closely related to a pangolin virus from Guangxi. Thus, it is very likely that just a single recombination event affected RaTG13, although there is a lot of uncertainty regarding the origin of the donor sequence.

The identification of multiple low-support recombination events that only differ in the departure point, which indicates the origin of the donor of the recombinant region, suggests that the signal of one recombination is diluted when using the ACGAnnotator tool that only summarizes recombinations that depart from the same branch. To classify these recombinations as one, a new summary ACG was obtained with our new implementation of this tool. Seven recombination events with a posterior probability support greater than 0.5 were recovered (Fig. 3) and as expected, one involved RaTG13. In this recombination, RaTG13 received most of the second half of the RBD from an unsampled virus that belongs to the same clade as the pangolin viruses from Guangdong and Guangxi, RaTG13, SARS-CoV-2, BANAL-103 and rshstt182. This recombination occurred 106 yBP (95% HPDI: 234 − 13 yBP) and the recombinant region, considering the HPD intervals of the start and end site (Table 1), includes the six contact amino acids of the variable loop. These results indicate that RaTg13 was indeed involved in a recombination event that likely resulted in the loss of five of the six contact amino acids.

Table 1

Median estimates of dates, and recombinant regions of the seven recombination events recovered by Bacter. The recombination numbers correspond to those in the text and Fig. 4. Key contact amino acids are located between positions 430–631.
Recombination number	Emergence time of donor sequence yBP (95% HPDI)	Recombination date yBP (95% HPDI)	Sites	Start site 95%HPD	End site 95% HPD
1	1021 (1728 − 493)	106 (234 − 13)	390–612	304–424	603–645
2	1370 (1804 − 735)	38 (98 − 10)	323–635	302–351	630–646
3	416 (666 − 194)	275 (536 − 35)	0–29	0–4	21–48
4	145 (329 − 61)	55 (122 − 14)	0–55	0–0	50–62
5	278 (412 − 148)	196 (375 − 25)	0-253	0-221	225–740
6	277 (412 − 143)	197 (376 − 26)	0-248	0-219	161–740
7	166 (332 − 53)	97 (241 − 14)	87–193	84–90	132–429

In recombination event #2, rs4084, a virus isolated from the Chinese rufous horseshoe bat (Rhinolophus sinicus), also acquired most of the second half of the RBD from an unsampled taxa related to Sarbecoviruses from Kenya and Bulgaria. While this recombination event is relatively recent, occurring 38 yBP (95% HPDI: 98 − 10 yBP), the lineage of the donor sequence originated over 1300 years ago. The host was likely an African or European bat species that eventually migrated to Asia. In recombination event #3, which happened 275 yBP (95% HPDI: 536 − 35 yBP), rshstt182, a virus isolated from the Shamel's horseshoe bat (Rhinolophus shameli) in Cambodia, received the first 29 nucleotides of the RBD from a virus closely related to rsyn09, which was isolated in China from the lesser brown horseshoe bat (Rhinolophus stheno).

The remaining recombination events involved Sarbecoviruses sampled in China. Recombination event #4 occurred 55 yBP (95% HPDI: 122 − 14 yBP). The point of origin was detected in the ancestral branch of two bat Sarbecoviruses infecting the greater horseshoe bat (Rhinolophus ferrumequinum) and the least horseshoe bat (Rhinolophus pusillus). The arrival point of the recombination was located in the ancestral branch of two Sarbecoviruses, also infecting the greater horseshoe bat. The recombinant region consisted of the first 55 sites of the RBD. The potential effects on viral fitness of the transfer of short recombinant regions at the beginning of the RBD (like in recombinations 3 and 4), if any, remains to be identified.

In recombination events #5 and #6 the first 250 sites of the RBD were transferred from two unsampled viruses related to a sequence isolated from the Chinese rufous horseshoe bat, to the ancestral branch of a virus isolated from the same bat species, although in a different province. These recombination events occurred almost simultaneously, 196 (95% HPDI: 412 − 148 yBP) and 197 yBP (95% HPDI: 412 − 143). Curiously, a descendant of the branch that received the nucleotides in recombinations 5 and 6, was in turn the donor sequence in recombination event #7. Slightly over 100 sites were transferred 97 yBP (95% HPDI: 241 − 14 yBP) to a Sarbecovirus infecting the least horseshoe bat.

In this study we conducted a recombination-aware phylogenetic analysis of the RBD region of 39 Sarbecoviruses. Multiple recombination events with a posterior probability support greater that 0.5 involving different Rhinolophus species were detected, which suggests a close interaction between bat populations. In this regard, R. sinicus, R. pusillus and R. affinis have overlapping geographical ranges, and the last two have been proposed as the likely host of SARS-CoV-2 progenitors³⁴. The presence of a significant recombination event within then RBD involving RaTG13, supports the common ancestor hypothesis, in which the bat virus lost all but one of the contact amino acids already present in the common ancestor of RaTg13, SARS-CoV-2, BANAL-103 and GD410721 (Fig. 4). This ancestral virus was likely a generalist pathogen capable of infecting different types of mammalian hosts, since laboratory studies have proved that SARS-CoV-2 can bind to the ACE2 receptors of cattle, cats and dogs³⁵. While no recombination events were detected between SARS-CoV-2 and BANAL-103, the fact that bat Sarbecoviruses, able to recognize human cell receptors, are found in nature, highlights the importance of broad sampling of bats and other potential host species of coronaviruses, as any newly discovered virus could improve our understanding of the origin of SARS-CoV-2 tremendously. Our results support a natural emergence (through vertical evolution combined with a number of recombination events) of the genetic region that makes SARS-CoV-2 so successful at transmitting among humans.

It is important to notice that only a small region of the genome was analyzed, and recombination breakpoints have been detected throughout the whole Sarbecovirus genome. Thus, recombination could still have been an important factor in the emergence of SARS-CoV-2, involving regions other than the variable loop of the RBD. In this regard, signals of recombination events involving the SARS-CoV-2 lineage have been detected on the 5’and 3’ ends of the S gene²⁷. In a recent study¹⁸, a recombination event was detected at the beginning of the SARS-CoV-2 RBD, associated with these viruses, however, it extended beyond the region included in our analysis. Recombination events involving other groups of coronaviruses could also be possible.

In this study we present a recombination-aware phylogenetic analysis of Sarbecoviruses, which sheds light on the evolutionary origin of SARS-CoV-2. Our simultaneous estimation of the vertical (tree-like) and horizontal (recombination) evolutionary history of the virus is in stark contrast to the more traditional approach that consists in the initial detection of recombination breakpoints followed by the phylogenetic reconstruction of each region located between breakpoints. While we recognize that the computational requirements of the employed approach restricted the scope of this study, as we couldn’t analyze the full data set and only analyzed a small fragment of the Sarbecovirus genome, we believe that the results obtained here provide an important “in-depth look” into the recombination history of the RBD. Further methodological developments could allow the analysis of full genes or even full genomes, which in turn will provide us with a more detailed overview of the role of recombination in the evolutionary history of coronaviruses and other pathogens.

Dataset

The dataset consisted of 87 genomes downloaded from the GenBank and GISAID databases (Supplementary Table S1 online). These included one SARS-CoV-2 sequence sampled in Wuhan in 2019, one SARS-CoV sequence, six Sarbecovirus sequences from pangolins, and 81 from bats. The sequences were aligned with MAFFT 7.475 using the default settings³⁶, producing a sequence alignment with 31208 sites. The program Gblocks²⁸ was then used to detect poorly aligned positions, which were replaced by Ns. A column of the alignment was masked if it contained at least 50% of gaps. Neighboring sites with more than 15% of gaps were also masked. The region composed by the RBD was extracted from the full genome alignment using the genomic locations of the SARS-CoV-2 reference genome NC_045512.2 to produce an RBD alignment with a length of 741 sites.

Substitution model and tree prior selection

The program bMoldeTest³⁰ implemented in the platform BEAST2³⁷ was used to select the best substitution model for the data set. The analysis ran for 200 million states. Next, a model comparison between the constant population size and the Bayesian skyline coalescent tree priors, was done with a Stepping-Stone sampling³². This comparison ran for 50 steps with a Markov Chain Monte Carlo (MCMC) length of 5000000 each, using the selected model of nucleotide substitutions. The Bayesian skyline coalescent was selected as the best model (logBF= 30.57).

Test for temporal signal

The evaluation of temporal signal for the alignment was done with the BETS approach²⁹. In this method, the marginal likelihoods of a heterochronous model, with the real dates (M1), and an isochronous model, with all dates set to zero and a fixed molecular clock rate (M2), are approximated in a Stepping-Stone sampling³². A positive log Bayes Factor value (log marginal likelihood M1 - log marginal likelihood M2) indicates the presence of temporal signal. The Path Sampling analysis ran for 50 steps with a Markov Chain Monte Carlo (MCMC) length of 5000000 each, using the best substitution model and tree prior.

Bayesian recombination analysis

The Bayesian recombination analysis of the RBD region was performed with Bacter, a BEAST2 package that generates a sample of Ancestral Recombination Graphs (ARGs) from the posterior distribution. An ARG consists of a backbone phylogeny, known as the Clonal Frame (C) that represents the true genealogy of the sequences, and a set of recombinant edges (R) that connect two branches of the tree and represent recombination events²⁴ (Fig. 2). A Bacter analysis can be computationally demanding, and initial analyses using the 87 sequences showed that it was not feasible to analyze the whole data set, given the time and amount of resources needed. Hence, a sample of the data set was generated with Uclust³³, and algorithm that generates clusters of sequences according to a certain threshold of genetic diversity. The cluster was selected using the cluster_fast command and a threshold of 96%, which represents the similarity between SARS-CoV-2 and RaTG13. 39 sequences were selected, including the SARS-CoV-2 genome and its closest relatives (RATG13, GD410721 and one cave bat virus).

In order to improve the mixing and convergence of the analyses, Bacter was used in combination with the Metropolis coupled Markov chain Monte Carlo algorithm (MC³). In this extension of the traditional MCMC, multiple parallel chains are used to sample the posterior. One, known as the cold chain explores the posterior like in the regular MCMC. The other runs are known as the “heated” chains and add an exponent β to the posterior probability. Values of β lower than one, reduce or “melt” the peaks of probability making easier the movement of the chains between regions of high probability. The cold and heated chains can exchange locations, reducing the risk of the analysis getting stuck at a local optimum³⁸. In BEAST2 the MC³method is implemented in the CoupledMCMC package³⁹.

Three independent analyses were done, each using 8 chains (7 heated) with a length of 500 million states. The Bayesian skyline tree prior was used with 5 population sizes, as well as the substitution model selected by bModelTest with the addition of 8 gamma rate categories to account for rate heterogeneity among sites. We also set an upper bound of 10 recombination events allowed per sampled ACG, as it is a practical solution to set an upper bound to reduce the run time of the analysis. For the prior distribution of the clock rate, a normal distribution (mean = 7.8E^-4, σ = 3.0E^-4) was used, similar to one previously employed in a molecular dating analysis of Sarbecoviruses¹¹. A normal distribution (mean = 150, σ = 50) was employed as a prior for the mean length of the recombinant part of the sequence. A uniform distribution between 0 and 100000 was used for the prior of the population size. Lastly a uniform distribution between 0 and 15000 years was also set as the prior for the age of the root of the clonal frame, considering the results of a previous dating analysis of Sarbecoviruses¹¹. The output files were combined with the software LogCombiner⁴⁰, after removing a 10% burn-in. This yielded effective sample sizes above the standard threshold of 200 for all parameters, as assessed by the software Tracer v.1.7⁴¹.

The maximum clade credibility tree of the clonal frame (without recombination events) was obtained for the combined runs using the program TreeAnnotator⁴² with a 10% burn-in. A summary ACG was obtained with the ACGAnnotator tool included in the Bacter package, using a posterior support threshold of 0.5 for the recombination events. The summary method employed by Bacter merges recombinations if they connect the same pair of branches and if the recombinant region overlaps at least by 1 base pair²⁴. However, there are cases where multiple recombinations arrive at the same branch but their origin points are in different locations of the clonal frame (e.g. just before or after an internal node). While Bacter considers these recombinations as independent events, it is possible that in reality they are the same recombinations, with the different origin points representing the uncertainty in the identity of the donor branch. To circumvent this problem, we also used a modified version of the ACGAnnotator code that merges recombinations with different origin points if they (i) arrive at the same branch and (ii) have overlapping recombinant regions.

Data availability

The authors declare that all data supporting the findings of this study are available within the paper, the Supplementary Information and in publicly available data repositories.

Multiple sequence alignment, xml, combined log and tree files are available at https://github.com/tidelab/Sarbecovirus_recombination. Note that 19 sequences fall under the terms of use of the GISAID platform.

The Bacter code with the new implementation of ACGAnnotaror tool is available at https://github.com/tidelab/bacter/tree/modify_summary_method

Acknowledgments

We would like to thank Prof. Dr. Johannes Krause, Dr. Kirsten Bos, Dr. Alexander Herbig and Dr. Sanni Översti for their valuable inputs and suggestions for the analyses.

Author contributions

LREG, AW, AK and DK participated in the design of the study. LREG, AW, AK and DK defined the parameter setup for the analyses. LREG performed the analyses. LREG wrote the draft of the manuscript. AW, AK and DK reviewed and edited the manuscript.

Supplementary Information

The online version contains supplementary information available at….

Competing interests

The authors declare no competing interests.

Li, Q. et al. Early Transmission Dynamics in Wuhan, China, of Novel Coronavirus–Infected Pneumonia. N. Engl. J. Med. 382, 1199–1207 (2020).
Zhou, P. et al. A pneumonia outbreak associated with a new coronavirus of probable bat origin. Nature 579, 270–273 (2020).
Gorbalenya, A. E. et al. The species Severe acute respiratory syndrome-related coronavirus: classifying 2019-nCoV and naming it SARS-CoV-2. Nat. Microbiol. 5, 536–544 (2020).
Chan-Yeung, M. & Xu, R. H. SARS: epidemiology. Respirology 8 Suppl, (2003).
Zaki, A. M., van Boheemen, S., Bestebroer, T. M., Osterhaus, A. D. M. E. & Fouchier, R. A. M. Isolation of a novel coronavirus from a man with pneumonia in Saudi Arabia. N. Engl. J. Med. 367, 1814–1820 (2012).
Mousavizadeh, L. & Ghasemi, S. Genotype and phenotype of COVID-19: Their roles in pathogenesis. J. Microbiol. Immunol. Infect. 0–4 (2020) doi:10.1016/j.jmii.2020.03.022.
Lan, J. et al. Structure of the SARS-CoV-2 spike receptor-binding domain bound to the ACE2 receptor. Nature 581, 215–220 (2020).
Lu, R. et al. Genomic characterisation and epidemiology of 2019 novel coronavirus: implications for virus origins and receptor binding. Lancet 395, 565–574 (2020).
Andersen, K. G., Rambaut, A., Lipkin, W. I., Holmes, E. C. & Garry, R. F. The proximal origin of SARS-CoV-2. Nat. Med. 26, 450–452 (2020).
Wan, Y., Shang, J., Graham, R., Baric, R. S. & Li, F. Receptor Recognition by the Novel Coronavirus from Wuhan: an Analysis Based on Decade-Long Structural Studies of SARS Coronavirus. J. Virol. 94, 1–9 (2020).
Boni, M. F. et al. Evolutionary origins of the SARS-CoV-2 sarbecovirus lineage responsible for the COVID-19 pandemic. Nat. Microbiol. 5, 1408–1417 (2020).
Zhou, H. et al. A Novel Bat Coronavirus Closely Related to SARS-CoV-2 Contains Natural Insertions at the S1/S2 Cleavage Site of the Spike Protein. Curr. Biol. 30, 2196–2203.e3 (2020).
Graham, R. L. & Baric, R. S. Recombination, reservoirs, and the modular spike: mechanisms of coronavirus cross-species transmission. J. Virol. 84, 3134–3146 (2010).
Xiao, K. et al. Isolation of SARS-CoV-2-related coronavirus from Malayan pangolins. Nature 583, 286–289 (2020).
Liu, P. et al. Are pangolins the intermediate host of the 2019 novel coronavirus (SARS-CoV-2)? PLoS Pathog. 16, 1–13 (2020).
Li, X. et al. Emergence of SARS-CoV-2 through recombination and strong purifying selection. Sci. Adv. 6, 1–14 (2020).
Tagliamonte, M. S. et al. Multiple recombination events and strong purifying selection at the origin of SARS-CoV-2 spike glycoprotein increased correlated dynamic movements. Int. J. Mol. Sci. 22, 1–16 (2021).
Temmam, S. et al. Coronaviruses with a SARS-CoV-2-like receptor- binding domain allowing ACE2-mediated entry into human cells isolated from bats of Indochinese peninsula. Res. Sq. (2021).
Lole, K. S. et al. Full-length human immunodeficiency virus type 1 genomes from subtype C-infected seroconverters in India, with evidence of intersubtype recombination. J. Virol. 73, 152–160 (1999).
Martin, D. P., Murrell, B., Golden, M., Khoosal, A. & Muhire, B. RDP4: Detection and analysis of recombination patterns in virus genomes. Virus Evol. 1, (2015).
Martin, D. P. et al. RDP5: a computer program for analyzing recombination in, and removing signals of recombination from, nucleotide sequence datasets. Virus Evol. 7, 87 (2021).
Pond, S. L. K., Posada, D., Gravenor, M. B., Woelk, C. H. & Frost, S. D. W. Automated phylogenetic detection of recombination using a genetic algorithm. Mol. Biol. Evol. 23, 1891–1901 (2006).
Rasmussen, M. D., Hubisz, M. J., Gronau, I. & Siepel, A. Genome-Wide Inference of Ancestral Recombination Graphs. PLOS Genet. 10, e1004342 (2014).
Vaughan, T. G. et al. Inferring ancestral recombination graphs from bacterial genomic data. Genetics 205, 857–870 (2017).
Hubisz, M. & Siepel, A. Inference of Ancestral Recombination Graphs Using ARGweaver. Methods Mol. Biol. 2090, 231–266 (2020).
Didelot, X., Lawson, D., Darling, A. & Falush, D. Inference of Homologous Recombination in Bacteria Using Whole-Genome Sequences. Genetics 186, 1435 (2010).
Müller, N. F., Kistler, K. E. & Bedford, T. A Bayesian approach to infer recombination patterns in coronaviruses. Nat. Commun. 2022 131 13, 1–9 (2022).
Castresana, J. Selection of Conserved Blocks from Multiple Alignments for Their Use in Phylogenetic Analysis. Mol. Biol. Evol. 17, 540–552 (2000).
Duchene, S. et al. Bayesian evaluation of temporal signal in measurably evolving populations. Mol. Biol. Evol. 37, 3363–3379 (2020).
Bouckaert, R. R. & Drummond, A. J. bModelTest: Bayesian phylogenetic site model averaging and model comparison. BMC Evol. Biol. 17, 1–11 (2017).
Hasegawa, M., Kishino, H. & Yano, T. aki. Dating of the human-ape splitting by a molecular clock of mitochondrial DNA. J. Mol. Evol. 22, 160–174 (1985).
Xie, W., Lewis, P. O., Fan, Y., Kuo, L. & Chen, M. H. Improving marginal likelihood estimation for Bayesian phylogenetic model selection. Syst. Biol. 60, 150–160 (2011).
Edgar, R. C. & Bateman, A. Search and clustering orders of magnitude faster than BLAST. Bioinformatics 26, 2460–2461 (2010).
Lytras, S. et al. Exploring the Natural Origins of SARS-CoV-2 in the Light of Recombination. Genome Biol. Evol. 14, (2022).
Conceicao, C. et al. The SARS-CoV-2 Spike protein has a broad tropism for mammalian ACE2 proteins. PLOS Biol. 18, e3001016 (2020).
Katoh, K. & Standley, D. M. MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Mol. Biol. Evol. 30, 772–780 (2013).
Bouckaert, R. et al. BEAST 2.5: An advanced software platform for Bayesian evolutionary analysis. PLoS Comput. Biol. 15, (2019).
Altekar, G., Dwarkadas, S., Huelsenbeck, J. P. & Ronquist, F. Parallel Metropolis coupled Markov chain Monte Carlo for Bayesian phylogenetic inference. Bioinformatics 20, 407–415 (2004).
Müller, N. F. & Bouckaert, R. R. Adaptive metropolis-coupled MCMC for BEAST 2. PeerJ 8, e9473 (2020).
LogCombiner | BEAST Documentation. https://beast.community/logcombiner.
Rambaut, A., Drummond, A. J., Xie, D., Baele, G. & Suchard, M. A. Posterior Summarization in Bayesian Phylogenetics Using Tracer 1.7. Syst. Biol. 67, 901–904 (2018).
BEAST 2. https://www.beast2.org/treeannotator/.

No competing interests reported.

Download PDF

Journal Publication

published 04 Jan, 2024

Read the published version in Scientific Reports →

Editorial decision: Major revision
11 Jan, 2023
Reviews received at journal
22 Dec, 2022
Reviewers agreed at journal
16 Dec, 2022
Reviews received at journal
08 Dec, 2022
Reviewers agreed at journal
08 Dec, 2022
Reviewers agreed at journal
06 Dec, 2022
Reviewers invited by journal
05 Dec, 2022
Editor assigned by journal
05 Dec, 2022
Editor invited by journal
18 Nov, 2022
Submission checks completed at journal
18 Nov, 2022
First submitted to journal
21 Oct, 2022

You are reading this latest preprint version

Recombination-aware phylogenetic analysis sheds light on the evolutionary origin of SARS-CoV-2

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Results

Temporal signal and model selection

Bacter analysis and dating of the clonal frame

Bacter analysis and recombination events

Discussion

Methods

Declarations

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1