Global lineage evolution pattern of sars-cov-2 in Africa, America, Europe, and Asia: A comparative analysis of variant clusters and their relevance across continents

Introduction

The emergence and rapid spread of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) have caused a global pandemic with devastating consequences for public health and the global economy.^[1] SARS-CoV-2 is a positive-sense, single-stranded RNA virus that belongs to the family Coronaviridae.^[2] It is the causative agent of coronavirus disease 2019 (COVID-19), which has affected millions of people worldwide.^{[3, 4]} SARS-CoV-2 is thought to have originated in bats and was transmitted to humans through an intermediate host, possibly a pangolin. Since its initial identification in December 2019, SARS-CoV-2 has undergone rapid evolution, with several distinct lineages emerging worldwide. These lineages differ in their genetic makeup and have been associated with differences in disease severity, transmissibility, and vaccine efficacy.^[5] Understanding the lineage evolution of SARS-CoV-2 is critical for developing effective strategies to control the pandemic and prevent future outbreaks.

SARS-CoV-2 is an RNA virus that has undergone various evolutionary changes since its initial emergence in late 2019. Several genetic studies have found that the virus has a relatively slow mutation rate compared to other RNA viruses, which may be due to a proofreading mechanism that corrects errors during replication.^[6] Despite its slow mutation rate, SARS-CoV-2 has still undergone significant evolutionary changes that have contributed to its spread and ability to evade the immune system.^[7]

Several factors have contributed to the emergence of new variants, including the high global prevalence of the virus, the large number of infected individuals, and the limited access to vaccines in many parts of the world.^[8] Additionally, the evolution of the virus is likely influenced by various host factors, such as individual immune responses and underlying health conditions.^[9] Efforts are underway to monitor and track the evolution of SARS-CoV-2, including through genomic surveillance programs that sequence the virus from infected individuals around the world.^[10] This information is used to track the emergence of new variants and inform public health strategies to control the spread of the virus.^[11]

Tumors are known to evolve over time as a result of genetic mutations and natural selection, leading to the emergence of increasingly aggressive and treatment-resistant cancer cells.^[12] Similarly, viruses such as SARS-CoV-2 can undergo rapid evolution through mutations and natural selection, allowing them to adapt to changing host environments and potentially evade immune responses.^[13] By applying the principles of tumor evolution to the study of viral evolution, we can gain a deeper understanding of how SARS-CoV-2 evolves over time. This can have important implications for the development of vaccines and antiviral therapies, as well as our understanding of viral pathogenesis and the emergence of novel viral strains.

Phylogenetic inference can be employed to reconstruct clonal lineages and to comprehend tumor evolution by utilizing sequencing data.^{[14, 15]} In the case of phylogenetic trees, the common ancestors are represented by internal nodes, and their genotype can be inferred based on the similarities between their descendants.^[16] Hence, phylogenetic trees offer a way to investigate the past by determining the sequence in which mutations occurred as clones diverged in lineages and formed subpopulations.

The same principles used to study tumor evolution through phylogenetic inference can also be applied to the evolution of viruses, such as the SARS-CoV-2 virus. Viral evolution is driven by the same basic principles of natural selection and genetic drift as other organisms, with mutations accumulating over time and being passed down to subsequent generations.^[17] In the case of SARS-CoV-2, the virus has undergone multiple rounds of mutation and selection since its initial emergence in late 2019, resulting in the emergence of numerous viral variants with different characteristics.

Phylogenetic inference can be used to reconstruct the evolutionary history of the SARS-CoV-2 virus and track the emergence and spread of different viral variants. By analyzing the genetic sequences of different viral isolates and constructing a phylogenetic tree, researchers can estimate the order in which mutations occurred and identify the common ancestors of different viral lineages.^[18] This can provide important insights into the origins and transmission patterns of different viral variants, as well as their potential impact on disease severity and transmissibility.

One challenge in studying viral evolution is the rapid pace at which mutations can occur, as well as the potential for convergent evolution, where different viral lineages independently evolve similar mutations in response to similar selective pressures. However, the use of phylogenetic inference, combined with other methods such as experimental studies and epidemiological analysis, can help to provide a more complete picture of how the SARS-CoV-2 virus is evolving over time and how it may continue to evolve in the future.

In this paper, we explored the parallels between tumor evolution and SARS-CoV-2 evolution, highlighting the common mechanisms of genetic variation, natural selection, and adaptation. We applied concepts from tumor evolution to the study of viral evolution, including the use of phylogenetic analysis to track the spread of SARS-CoV-2 variants.

We analyzed the spread and evolution of 74, 075 SARS-CoV-2 sequencing data from multiple regions worldwide, including Africa, America, Europe, and Asia. The mutations accumulated over an extended period were analyzed sequentially through the tumor evolution method to reconstruct the temporal order of mutations. Moreover, we reconstructed groups of mutations that propagate through regions with locally similar mutations. This analysis shows that the virus evolution process shares similarities with the lineage evolution process observed in tumors. Our findings suggest that studying SARS-CoV-2 through the lens of tumor evolution can provide a better understanding of its evolutionary dynamics.

Material and method

Result

Global evolutionary landscape of the SARS-CoV-2 virus

From GISAID, we assembled and analyzed 32 SARS-CoV-2 virus phylogenetic lineage trees from 18 African, 6 American, 4 European, and 4 Asian regions (Figure 1).

Figure 1

SARS-CoV-2 evolutionary lineages: (A) 18 nations or areas in Africa; (B) 6 nations or areas in the Americas; (C) 4 nations or areas in Europe; and (D) 4 nations or areas in Asia. Data regarding these lineages was sourced from GISAID. Classification of SARS-CoV-2 clades, referred GRA to O, is presented in Table 1. (E) Schematic lineage showed the three distinct patterns of virus evolution that arise throughout the progression of SARS-CoV-2.

Each tree ultimately culminated in omicron becoming the dominant species, and the patterns of evolution of the lineages were largely similar across regions in global view. In particular, the patterns of evolution of the lineages were similar to traditional tumor evolutionary lineage trees (Figure 1A-D). However, when examined over time, while tumor lineages survive by maintaining heterogeneity in the long term within tumor, in the case of the SARS-CoV-2 virus, each lineage disappeared in the final stages of the virus lineage, with new species emerging from other lineages.

Given the isolation of each region, it seems likely that the process of mutation between regions and even continents and the global pattern of lineage evolution in each region and the timing of divergence are similar. Nevertheless, we were able to identify the possibility of regionally independent variation. In particular, we found regions with atypical evolutionary trees, such as Senegal, West Africa, and Denmark (Figure 1A and C). They show an atypical pattern when analyzed with the available samples of the SARS-CoV-2 virus. The other regions are generally similar, which is noteworthy given the potential for a pandemic at the global level.

When analyzing their global patterns based on the clade classification of SARS-CoV-2 virus presented in Table 1, evolutionary patterns of the SARS-CoV-2 virus emerge from the perspective of cancer evolution, we found similar trends have been observed in the SARS-CoV-2 virus evolution.^{[16, 19]} For example, G, GH, GR, GRY proceeded with neutral evolution while GK generally showed a typical branching evolution. In particular, GRA (omicron), which appears to be typical of punctuated evolution. Punctuated tumor evolution is a pattern of clonal expansion in which early mutations survive their clonal sweep by slowly accumulating mutations. In the case of omicrons, we can see that many sub-lineage SARS-CoV-2 viruses accumulate mutations that are initially undetectable, slowly accumulating variation, and then suddenly exploding in population to become the dominant species in pandemic. (Figure 1E). It can be inferred that they have relatively many shared mutations and have survived clonal sweeping for a long period of time to reach their final state.

Table 1

Clade classification of SARS-Cov-2^[20]

Clade	Context of marker variants
	C241T, C3037T, A23403G, G28882A includes S: D614G+N: G204R+(at least 6 of the following amino acid
GRA (Omicron)	changes S: V143del, Y145del, N211del, ins214EPE, G339D, S371L, S477N, T478K, E484A, Q493R, Q498R, T547K, N679K, P681H, P681R, N764K, D796Y, N856K, Q954H, N969K, L981F)
GRY	C241T, C3037T, 21765: 21770del, 21991: 21993del, A23063T, A23403G, G28882A includes S: H69del, V70del, Y144del, N501Y+D614G+N: G204R
GH	C241T, C3037T, A23403G, G25563T includes S: D614G+NS3: Q57H
GR	C241T, C3037T, A23403G, G28882A includes S: D614G+N: G204R
G	C241T, C3037T, A23403G includes S: D614G
GK	C241T, C3037T, A23403G, C22995A includes S: D614G+T478K
GV	C241T, C3037T, A23403G, C22227T includes S: D614G+A222V
S	C8782T, T28144C includes NS8: L84S
L	C241, C3037, A23403, C8782, G11083, G26144, T28144(early clade markers in WIV04: reference sequence)
V	G11083T, G26144T NSP6: L37F+NS3: G251V
O*	-

1. *Others.

Global lineage evolution pattern of Africa

We conducted an analysis of the history of SARS-CoV-2 variants that occurred in different countries and regions, focusing on the evolution of the omicron variant. The punctuated evolution pattern of omicron, characterized by the accumulation of mutations over a relatively long period of time, suggests that clonal sweeping may have occurred. In tumor evolution, the ancestral lineage of shared mutations is often obscured by clonal sweeping, but we sought to identify evolutionary patterns in SARS-CoV-2 by comparing mutations across regions.

By integrating data from 18 African countries, we were able to infer the order of shared mutations from the earliest to the latest branching mutations, as shown in Figure 2A and A’. Through an integrated analysis of the degree of mutation sharing with other countries and globally, we identified eight clusters based on their interconnectivity. (Table 2) In the case of the red C1 cluster, the earliest mutations were observed in all African countries, and as the evolutionary march progressed, there were mutations exclusively found in each country.

Figure 2

Clustering of 77 GRA variants and their pairwise Pearson’s correlation following hierarchical weight; (A) The correlation of SARS-CoV-2 variants in 18 distinct countries or regions of Africa. (B-D) Based on the relevance of SARS-CoV-2 lineages among African countries, the countries or regions were categorized into three distinct groups (Figure 4C). Pearson’s correlation analysis was performed using seaborn in Python and applied to hCoV-19 phylogenetic data available on GISAID. (A'-D') Reconstructed phylogenetic lineage trees based on variant correlations and hierarchical evaluation scores in Africa (Figure S1A). Dashed branches indicate lineages for which bifurcation discrimination is not possible. Correlated variants, manually sorted according to Pearson’s correlation based on worldwide variants information, exhibited uniform color-coding throughout all plots. Variants reported only once were categorized as private. (E) The countries classified into three groups were geographically mapped.

Table 2

Clusters according to Pairwise Pearson’s correlation

Cluster	Mutation list
C1	S371L, N501Y, Q493R, S477N, Q498R, E484A, Y505H, G339D, N440K, S375F, S373P, T9I, A63T, Q954H, N969K, D796Y, N764K, I5967V, P3395H, P13L, H655Y, N679K, Q19E, K417N, S371F, P681H, G142D, T3255I, T478K
C2	D61L
C3	L3201F, T223I, T376A, D405N, R408S, L3027F, S413R, R5716C, T842I, S135R, G1307S, T19I, V213G, T6564I, T3090I
C4	F486V, D3N, L452R, P26Q, A27S
C5	L371F, F3677L, R856K
C6	G3676S, N211I
C7	Y145D, G446S, L2084I, F371L, K856R, D3G, R346K, G496S, A67V, L3674F, A2710T, T95I, I3758V, T547K, N856K, L981F
C8	Y144*, V70I
private*	R204G, S391F, S3675F, T2710A, I95T, D145Y, A2909V, P25T, S446G, V3758I, G3D, L212I, L106F, D343G, V143H, S2083N, L212V, T3646A, L2084V, S2083T, K440N, H69V, Y144L, G204R, S33G, R493Q

1. *Mutation that occurred only once in all phylogenetic trees.

Our analysis revealed three distinct patterns of variant occurrence within the African continent, which we designated as group 1 (Algeria, Angola, Burundi, Ethiopia, Niger), group 2 (Central African Republic, Congo-Brazzaville, Congo-Kinshasa, Uganda), and group 3 (Botswana, Cameron, Gabon, Madagascar, Malawi, Nigeria), as shown in Figure 2B-D. These groups were geographically distributed from north to south (Figure 2E), and shared the same initial variant occurrence, characterized by mutations in E: T9I, M: Q19E, A63T, S: G339D, S373P, S375F, and orf1ab: P3395H (Figure 2B’-D’). However, groups 1 and 3 were trend-separated, with group 2 bridging the trend in the middle (Figure 4C).

Figure 3

(A-C) Clustering of 92 GRA variants and their pairwise Pearson’s correlation following hierarchical weight. (A'-C') Reconstructed phylogenetic lineage trees based on variant correlations and hierarchical evaluation scores in each continent (Figure S1A). Dashed branches indicate lineages for which bifurcation discrimination is not possible. Correlated variants, manually sorted according to Pearson’s correlation based on worldwide variants information, exhibited uniform color-coding throughout all plots. Variants reported only once were categorized as private.

Figure 4

(A) Clustering of 99 GRA variants and their pairwise Pearson’s correlation following hierarchical weight and (A') reconstructed phylogenetic lineage tree on the basis of variant correlations and hierarchical evaluation scores in Worldwide (Figure S1B). Variants clustered on Pearson’s correlation were each indicated by different color. Variants that reported only a single time were labeled as private. The underline indicates a pre-branching mutation commonly founded across all continents. Clustering countries or regions according to relevance of hierarchical weight of GRA phylogenetic lineages; (B) 32 Worldwide; 18 Africa; (D) 4 Asia, 4 Europe and 6 America. According to the pairwise Pearson’s correlation, the three distinct groups were classified as Group 1-3. (E) Dimensionality reduction analysis visualized. The dataset used for executing all three dimensionality reduction approaches was derived from the SARS-CoV-2 phylogenetic information accessible on GISAID. The input data incorporated 99 GRA variants along with 32 phylogenies information originating from distinct countries or regions (Figure S1B). Variants, manually sorted according to Pearson’s correlation coefficient, exhibited uniform color-coding throughout all four plots.

The analysis suggests that even within the relatively large African continent, there is evidence of shared patterns of variant occurrence, possibly influenced by geographic proximity. Furthermore, the identification of shared mutations across regions may provide insight into the ancestry of SARS-CoV-2 variants, despite the occurrence of clonal sweeping.

Discussion

We conducted a comprehensive analysis of the SARS-CoV-2 virus phylogenetic lineage trees derived from 18 African, 6 American, 4 European, and 4 Asian regions using data from GISAID. Our analysis revealed that the patterns of evolution of the lineages leading to the emergence of omicron, the currently dominant species, were largely similar across regions in global view. However, it is clear that the virus evolved differently across different continents, with each region exhibiting its unique variant clusters and patterns while maintaining the global interaction leading to similar marching evolution. The integration of the results from each continent allowed us to identify eight correlated variant clusters, and the worldwide lineage tree was reconstructed based on the occurrence of variants and their relatedness.

The method of assigning weights based on the hierarchical positions of variants allowed for the evaluation of the global importance of each variant and its role in the evolution of SARS-CoV-2. It also enabled the identification of variants that were highly conserved across multiple regions or countries, as well as those that were unique to specific regions. The use of exponential weights also provided a more nuanced analysis of the contribution of each variant to the evolution of the virus, compared to a simple binary classification of the presence or absence of a variant in a given phylogenetic lineage tree.

These patterns of evolution were observed to be similar to those observed in traditional tumor evolutionary lineage trees,^{[16, 21, 22]} as shown in Figure 1. However, the SARSCoV-2 strains appeared to have diverged from the viral lineage in the final stages, unlike tumor lineages that maintain heterogeneity in the long term for survival. There are many reasons why a particular SARS-CoV-2 virus lineage may disappear, including the development of a vaccine or a treatment that works better for certain lineages or a lethal event that prevents further progression of infection.^{[7, 23, 24]} Unlike tumor clonal sweeping,^{[25, 26, 27]} in the case of the SARS-CoV-2 virus, a particular lineage may explode in infection at a particular time point and then suddenly disappear.^[28] It is not clear whether this is the case for tumors, but in cases with strong germline drivers in early neonatal cancer,^{[29, 30]} there is chance for each cancer lineage to become many different evolutionary forms of cancer in neutral or branching patterns.^[31] We did not observe the SARS-CoV-2 virus where a single lineage reemerged as a new pattern of lineages in the long term. This suggests that further research is needed to determine if there are any patterns of SARS-CoV-2 virus that can survive in the long term without causing the severe pathological event rather, it is a form of infectivity that survives in an individual or inter-individual for a long period of time and remains infectious.

Our analysis is essentially based on the similarity between the characteristics of a tumor that has undergone many different sweeps over time and the characteristics of a long-term latent SARS-CoV-2 virus. We hypothesized that this could potentially allow for attempts to explain the evolutionary pattern of SARS-CoV-2, presenting a new perspective for analyzing the virus through the lens of cancer evolution. In particular, the Omicron variant, a new dominant strain that appeared after the extinction of the Delta variant, was initially undetected for a long period and then reported first in Africa along with numerous accumulated mutations.^{[32, 33]} This phenomenon closely mirrors the punctuated lineage model of cancer evolution. ^[16] Based on these aspects, it was possible to metaphorically view the emergence and extinction of numerous SARS-CoV-2 variants, driven by factors such as natural selection, policy interventions, and vaccination, in parallel with the process of potential cancer cells being extinguished by the immune system or relapsing after chemotherapy due to dormant cancer stem cells. A commonality between these two phenomena was that entities not detected through any channels and survived for a long time, accumulating mutations, eventually become dominant (Figure 1). Notably, Omicron has a “long branch”, which can also be called “lost time period”, where mutations accumulated over a long period until being reported worldwide, making it impossible to know the order of occurrence. By integrating phylogenetic information from various regions or continents about this, it was possible to estimate the order of mutation occurrence (Figure 4A’).

Based on the analysis of the global lineage evolution pattern of SARS-CoV-2, it is clear that the virus evolved differently across different continents, with each region exhibiting its unique variant clusters and patterns. The integration of the results from each continent allowed the researchers to identify eight correlated variant clusters, and the worldwide lineage tree was reconstructed based on the occurrence of variants and their relatedness. The reconstructed tree revealed that once all the variants in Cluster 1 occurred, it diverged into two lineages based on the occurrence of specific variants.

The differences in evolutionary patterns between various regions or continents were identified in this article. Since our study only looked at variants according to the WGS of viruses, we do have the limitation on it. It seems reasonable to consider not only viral factors but also host factors to explain them. In this context, the similar distribution of branches in a large pandemic can be seen as an indication of the impact of multiple and non-isolated global communications. Ultimately, however, the difference is likely to be a function of the local, which has much more interaction with each other than the global level of communication. First, vaccination not only prevents the spread of the virus but can also induce a genetic bottleneck. Therefore, it is possible that the patterns between regions where vaccination has been successfully administered and those where it hasn’t slightly show the different patterns of the viral evolution. Particularly, the pace and ratio of vaccination can be highly linked to socioeconomic factors, with economically affluent countries having a higher vaccination rate than those less affluent. On this aspect, several research was reported that high and low vaccine coverage may affect the pattern of the viral evolution based on the conducted simulation or surveillance under such assumptions.^{[34, 35, 36]} Under this background, further investigation is required for a better understanding of the correlation of artificial interventions with evolutionary patterns. Second, it seems to have a significant influence in terms of the viral evolution, such as “the order of isolation.” Strong isolation and quarantine in particular are known to have different initial spread effects that we were unable to capture in our study.^[37]

We also acknowledge the possibility that the mutation process could have been influenced by transportation by airplanes, vehicles, etc.^{[38, 39]} Despite the isolation of each region, we observed that the global pattern of lineage evolution in each region was similar, and the timing of divergence was also similar, indicating a possible interregional influence. Nevertheless, our analysis identified regions with atypical evolutionary trees, such as Senegal, West Africa, and Denmark, which could be potentially due to overall timing or sample bias. Independent pathways and diverging lineages were observed, indicating the complex and dynamic nature of SARS-CoV-2 evolution. Interestingly, the order of variant occurrence within clusters was different across continents, with Africa having a different intra-cluster order than the other three continents. Furthermore, while Asia had a relatively isolated evolutionary lineage, the other continents shared many common evolutionary pathways.

These findings have important implications for understanding the global spread and evolution of SARS-CoV-2. The unique variant clusters and patterns across continents suggest that certain regions may have contributed more to the evolution and spread of the virus than others. This knowledge could be useful in developing targeted interventions to control the spread of the virus in specific regions. Additionally, the identification of independent pathways and diverging lineages highlights the importance of continued genomic surveillance and sequencing efforts to monitor the evolution of SARS-CoV-2 and identify new variants of concern. This knowledge can inform the development of effective vaccines and therapeutics to combat the ongoing pandemic and potential future outbreaks.^{[40, 41]}

Our findings suggest that similar to cancer evolution, some evolutionary patterns emerge in the global evolution of SARS-CoV-2 virus. For instance, certain lineages, such as G, GH, GR, GRY, showed neutral or branching evolution, while GK demonstrated typical branching evolution. Furthermore, GRA (omicron), which appears to be typical of punctuated evolution, is noteworthy.^{[42, 43]} Punctuated tumor evolution is characterized by clonal expansion, where early mutations survive clonal sweep by slowly accumulating mutations. In the case of omicrons, many sublineage SARS-CoV-2 viruses accumulated initially undetectable mutations, gradually accumulating variation before suddenly exploding in population to become the dominant species. The presence of relatively many shared mutations indicates that they have survived clonal sweeping for an extended period before reaching their final state. Overall, this interdisciplinary approach to studying viral pathogens through the lens of tumor evolution has the potential to offer new insights into the emergence and spread of viral pathogens, with significant implications for public health and clinical practice.

As we proceeded with our analysis, we observed that certain mutations shared ancestry at different points in time in continents with different hierarchies. For example, we observed mutations like S: S371L, P681H, and orf1ab: R5716C that seemed to form independent pathways and then rejoined in the offspring lineage, and S: A67V, orf1ab: L3674F that diverged and then rejoined in the offspring lineage. Such phenomena are relatively infrequent in human cancer, likely due to the larger genome size of humans. However, in the case of SARS-CoV-2, the relatively small size of its genome may facilitate such occurrences. Nonetheless, the reappearance or repair of specific mutations under certain conditions may indicate a structural tendency of the virus to be mutation-prone, or the existence of conditions that favor their occurrence. Further investigation is required to identify mutation hotspots in these viruses.

In conclusion, the global lineage evolution pattern of SARS-CoV-2 is complex and dynamic, with different regions exhibiting unique clusters and patterns of variants. The integration of results from multiple continents provides a more comprehensive understanding of the evolution of the virus and can inform targeted interventions and genomic surveillance efforts. Overall, these findings have important implications for public health and clinical practice. By studying the evolutionary patterns of SARS-CoV-2, we can gain insight into the emergence and spread of viral pathogens, as well as the development of resistance to antiviral treatments. Additionally, the similarities between tumor evolution and SARS-CoV-2 evolution suggest that new therapeutic approaches developed for cancer may also be applicable to viral diseases.